2026-01-28_abc_project_weekly_planning_68fec8a5

Meeting Title: ABC Project Weekly Planning Date: 2026-01-28 Meeting participants: Casie Aviles, Samuel Roberts, Amber Lin, Mustafa Raja

---

WEBVTT

1 00:00:41.860 ⇒ 00:00:42.920 Samuel Roberts: Okay.

2 00:00:44.470 ⇒ 00:00:45.210 Casie Aviles: Hey, son.

3 00:00:48.800 ⇒ 00:00:49.949 Samuel Roberts: How’s today going?

4 00:00:53.310 ⇒ 00:00:56.990 Casie Aviles: Good so far, I’m just, working on ABC for now.

5 00:00:57.810 ⇒ 00:00:58.460 Samuel Roberts: Cool.

6 00:01:31.640 ⇒ 00:01:32.690 Amber Lin: Hello!

7 00:01:34.500 ⇒ 00:01:35.290 Samuel Roberts: Hello.

8 00:01:38.030 ⇒ 00:01:38.959 Samuel Roberts: How are you going? Okay.

9 00:01:39.390 ⇒ 00:01:47.350 Amber Lin: I’m good. Let’s see. I think on my plate, I wanted to check, one, on how…

10 00:01:47.460 ⇒ 00:01:55.140 Amber Lin: our progress is based on the Gantt chart. If we’re behind, if we need our adjustments, how, if we’re…

11 00:01:55.140 ⇒ 00:02:11.909 Amber Lin: expecting to spend more time on certain things, or spending more time than we expected. And then two, like, I want to… want us to run a checklist of each department, of how they are, where we think they are at, so I can give that to the client, because they’ve been… they’ve been wanting to know, like, what… where we’re at.

12 00:02:12.230 ⇒ 00:02:13.720 Amber Lin: for each department.

13 00:02:15.650 ⇒ 00:02:17.760 Samuel Roberts: When you say each department, you mean…

14 00:02:17.760 ⇒ 00:02:21.890 Amber Lin: Like, pests, mechanical, lawn…

15 00:02:21.890 ⇒ 00:02:24.570 Samuel Roberts: Specific stuff, like the ZIPS database stuff, or…

16 00:02:24.890 ⇒ 00:02:43.239 Amber Lin: Yeah, so that will include, one, are all the zips in there? Are they accurate? Because I know, we’re… I think we’re missing some stuff on the pest commercial stuff, and missing a little bit of stuff on the mechanical. And then also, for the essential docks, we probably should do an audit of how their central.

17 00:02:43.240 ⇒ 00:02:43.560 Samuel Roberts: Yeah.

18 00:02:43.560 ⇒ 00:02:58.969 Amber Lin: Central Doc is, and what we recommend that they change, so that, like, they can do the work instead of they coming to us, and then it’s gonna be our job to help them fix it. But if we tell them early enough, like, hey, this is what we need you to do, then, like, it’s less work for us.

19 00:03:00.830 ⇒ 00:03:01.560 Samuel Roberts: Cool.

20 00:03:01.910 ⇒ 00:03:04.590 Amber Lin: Yeah. Okay.

21 00:03:05.120 ⇒ 00:03:11.609 Amber Lin: Do we want to pull up the Gantt chart, and then we can go through it to see where we’re at?

22 00:03:14.080 ⇒ 00:03:15.619 Amber Lin: Yeah, that’s probably a good idea.

23 00:03:16.180 ⇒ 00:03:17.110 Amber Lin: Okay.

24 00:03:22.020 ⇒ 00:03:25.610 Amber Lin: Alright, I have… I have the Gantt chart pulled up.

25 00:03:26.120 ⇒ 00:03:33.650 Amber Lin: So, let’s see… Let’s see here, where are we at right now?

26 00:03:36.570 ⇒ 00:03:38.880 Casie Aviles: We moved it… we moved it around.

27 00:03:39.430 ⇒ 00:03:41.859 Casie Aviles: We pushed some things back.

28 00:03:42.430 ⇒ 00:03:43.789 Amber Lin: Hmm, I see.

29 00:03:44.710 ⇒ 00:03:49.390 Casie Aviles: Right now… At least for… for me, I’m working on…

30 00:03:50.370 ⇒ 00:03:53.959 Casie Aviles: See the Implement Master Entry Endpoint.

31 00:03:54.860 ⇒ 00:03:56.710 Casie Aviles: So that’s number 10.

32 00:03:57.360 ⇒ 00:04:00.350 Casie Aviles: And then I also have…

33 00:04:00.970 ⇒ 00:04:06.100 Casie Aviles: So that one, I can take on that. I marked that as higher priority.

34 00:04:07.110 ⇒ 00:04:08.080 Casie Aviles: This one?

35 00:04:08.080 ⇒ 00:04:09.480 Amber Lin: November 11th?

36 00:04:10.370 ⇒ 00:04:11.830 Casie Aviles: No, number 10.

37 00:04:12.240 ⇒ 00:04:13.039 Amber Lin: Number 10.

38 00:04:13.770 ⇒ 00:04:17.109 Casie Aviles: So, 11 would… behalf.

39 00:04:17.110 ⇒ 00:04:18.800 Amber Lin: Like, 10 would be 5.

40 00:04:18.930 ⇒ 00:04:21.669 Casie Aviles: Yeah, I would say so, and then…

41 00:04:21.670 ⇒ 00:04:29.519 Amber Lin: 11 would be… Like, lower, around 3, or 1. Education.

42 00:04:29.520 ⇒ 00:04:30.130 Casie Aviles: longer.

43 00:04:30.540 ⇒ 00:04:31.399 Amber Lin: That’s fantastic. I see.

44 00:04:31.400 ⇒ 00:04:33.689 Casie Aviles: take longer, and it’s more like…

45 00:04:33.690 ⇒ 00:04:34.430 Amber Lin: Okay.

46 00:04:35.060 ⇒ 00:04:43.229 Casie Aviles: Yeah, and then I also have the text-to-SQL… Accuracy, but that one…

47 00:04:43.760 ⇒ 00:04:50.019 Casie Aviles: It might… we might need to do a spike first, because we don’t know how exactly we’re gonna…

48 00:04:50.210 ⇒ 00:04:52.700 Casie Aviles: What’s the best way to validate?

49 00:04:52.890 ⇒ 00:04:57.579 Casie Aviles: the text to SQL, since the data that we have is, you know, constantly changing.

50 00:04:58.400 ⇒ 00:04:59.350 Amber Lin: I see.

51 00:04:59.820 ⇒ 00:05:02.200 Amber Lin: I see, okay.

52 00:05:02.200 ⇒ 00:05:05.309 Samuel Roberts: I thought we talked about, using the dev, like.

53 00:05:05.950 ⇒ 00:05:08.369 Samuel Roberts: Like, a snapshot of the database, basically.

54 00:05:09.240 ⇒ 00:05:09.820 Casie Aviles: Yeah.

55 00:05:11.390 ⇒ 00:05:14.909 Samuel Roberts: So, like, we can just kind of, like, freeze it in time and test against that, maybe?

56 00:05:18.330 ⇒ 00:05:18.710 Casie Aviles: Okay.

57 00:05:18.710 ⇒ 00:05:19.280 Samuel Roberts: Okay.

58 00:05:20.240 ⇒ 00:05:24.830 Samuel Roberts: Like, the… the… especially, like, you have the dev databases set up for the admin UI.

59 00:05:25.260 ⇒ 00:05:26.399 Samuel Roberts: I’m wondering if we could, like, we don’t.

60 00:05:26.400 ⇒ 00:05:27.030 Amber Lin: Hmm.

61 00:05:27.440 ⇒ 00:05:33.399 Samuel Roberts: We want to know that it’s generating the right SQL, not necessarily that the right data’s coming through, so if we just test against that, that’s fine.

62 00:05:34.300 ⇒ 00:05:35.750 Casie Aviles: Oh, okay, okay.

63 00:05:36.210 ⇒ 00:05:51.769 Amber Lin: And we can also… I think a large part of the text-to-SQL error is coming from the way they ask the questions. So, like, we probably should go back in… into our logs to see what… how people ask their questions, because they asked, like.

64 00:05:51.770 ⇒ 00:05:57.699 Amber Lin: like, sometimes I’ll see their questions, like, that’s so weird, I don’t even know what they’re asking for. So…

65 00:05:57.700 ⇒ 00:06:03.349 Amber Lin: So, like, especially if we add the mechanical stuff, so, probably should do that as well.

66 00:06:03.900 ⇒ 00:06:15.840 Amber Lin: Cool, okay. Question on… on these, that’s currently… looks like they’re… we’re, like, in the middle of it, but are they in progress? Are they… like, how… was.

67 00:06:15.840 ⇒ 00:06:20.410 Samuel Roberts: What’s the status for these. Yeah, Mustafa can probably address that.

68 00:06:20.410 ⇒ 00:06:39.620 Mustafa Raja: Yeah, so, for implementing evals in Mastra, I was able to do that, yesterday. So, now we have a, Postgres storage in Mastra, where we, we, we, we are, capturing everything that is going on with our instance.

69 00:06:39.780 ⇒ 00:06:41.280 Mustafa Raja: And then…

70 00:06:41.840 ⇒ 00:06:47.869 Mustafa Raja: We do have scores, I guess the next step would be, you know, once we…

71 00:06:48.080 ⇒ 00:06:52.659 Mustafa Raja: the next step would be to, you know, push this stuff in… what’s it called, BigQuery? What do you say?

72 00:06:52.800 ⇒ 00:06:53.880 Mustafa Raja: Is that correct?

73 00:06:54.240 ⇒ 00:06:55.089 Mustafa Raja: Sam?

74 00:06:56.090 ⇒ 00:06:59.709 Samuel Roberts: Yeah, I think we also saw that there’s a Postgres we can do there too, right?

75 00:06:59.710 ⇒ 00:07:01.440 Mustafa Raja: Yeah, yeah, I did implement that.

76 00:07:02.120 ⇒ 00:07:05.099 Samuel Roberts: Right, but I mean, even on Google, does it need to be BigQuery?

77 00:07:05.360 ⇒ 00:07:10.620 Mustafa Raja: So BigQuery is going to be a replacement for Snowflake, right?

78 00:07:11.640 ⇒ 00:07:12.330 Casie Aviles: Hmm.

79 00:07:12.330 ⇒ 00:07:16.810 Samuel Roberts: Right, but, Monster didn’t have an easy connector for BigQuery, did it?

80 00:07:16.860 ⇒ 00:07:21.639 Mustafa Raja: No, they do not. So my idea is that we can…

81 00:07:21.750 ⇒ 00:07:25.039 Mustafa Raja: do a scheduled job on the Postgres.

82 00:07:25.170 ⇒ 00:07:26.130 Mustafa Raja: to convert…

83 00:07:26.130 ⇒ 00:07:26.810 Samuel Roberts: Oh, okay.

84 00:07:26.810 ⇒ 00:07:30.440 Mustafa Raja: the raw data, and push that to BigQuery.

85 00:07:31.200 ⇒ 00:07:32.779 Samuel Roberts: Okay, yeah, I think that works then.

86 00:07:33.260 ⇒ 00:07:45.359 Mustafa Raja: Yeah, because, what we… what we get, from, Mastra is mostly going to be, JSON columns, so we don’t need to, you know, run some job.

87 00:07:45.360 ⇒ 00:07:46.009 Samuel Roberts: That makes sense.

88 00:07:46.010 ⇒ 00:07:47.110 Mustafa Raja: inform that.

89 00:07:49.950 ⇒ 00:07:54.569 Mustafa Raja: I could share a sna- I could share the data… let me see if I can share the database with you.

90 00:07:59.220 ⇒ 00:08:14.750 Amber Lin: Cool. What about the other stuff? Like, for these 5 tasks that we have here, do you think it’s scheduled to… are we scheduled to finish it this Friday? Like, we have… we have it here ending on Friday, but I don’t know, like.

91 00:08:15.020 ⇒ 00:08:19.179 Amber Lin: Are we able to do it? And if… do we need to push it back?

92 00:08:19.740 ⇒ 00:08:23.000 Mustafa Raja: I think we need to push it back, I think for Slack… Yeah, all of them?

93 00:08:23.390 ⇒ 00:08:31.390 Mustafa Raja: Sorry, I think for Slack alerting, this cannot be set up until we have it on, what’s it called?

94 00:08:32.750 ⇒ 00:08:34.920 Mustafa Raja: On Cloud Run, I believe?

95 00:08:36.130 ⇒ 00:08:36.750 Mustafa Raja: Right?

96 00:08:37.760 ⇒ 00:08:39.200 Amber Lin: So we need, like, this.

97 00:08:40.590 ⇒ 00:08:42.030 Amber Lin: I’m a, America.

98 00:08:42.900 ⇒ 00:08:44.489 Samuel Roberts: Yeah, I think you’re probably right.

99 00:08:44.790 ⇒ 00:08:48.530 Samuel Roberts: I mean, we could set up, like, testing, but it’s not gonna be… you know.

100 00:08:48.820 ⇒ 00:08:49.630 Samuel Roberts: Sweet.

101 00:08:50.030 ⇒ 00:08:51.940 Mustafa Raja: I see.

102 00:08:51.940 ⇒ 00:08:52.320 Amber Lin: Hmm.

103 00:08:52.320 ⇒ 00:08:53.090 Mustafa Raja: Around, right?

104 00:08:53.090 ⇒ 00:08:54.120 Samuel Roberts: Yeah, I’d have to move that over.

105 00:08:54.120 ⇒ 00:08:57.750 Amber Lin: Here, so that kind of would be… Oh.

106 00:08:57.750 ⇒ 00:09:05.440 Mustafa Raja: And then, I believe the validate alerts we have forced something is just, a dependency, or 35, yeah.

107 00:09:05.690 ⇒ 00:09:07.029 Mustafa Raja: Sorry, 36 here.

108 00:09:07.270 ⇒ 00:09:10.199 Amber Lin: I see, so this would go, like, after that.

109 00:09:10.200 ⇒ 00:09:11.080 Mustafa Raja: Yeah. 2.

110 00:09:11.990 ⇒ 00:09:24.189 Amber Lin: I see, okay. So I’m moving this into, like, later February, so we have… for latency threshold hooks, like, is that gonna be done this Friday?

111 00:09:24.190 ⇒ 00:09:27.129 Mustafa Raja: Yeah, let’s talk about that. What do we really mean by that?

112 00:09:27.770 ⇒ 00:09:28.560 Amber Lin: I don’t know.

113 00:09:29.190 ⇒ 00:09:33.069 Samuel Roberts: I think that’s what’s going into the Slack alerts, so that if there’s, like, big…

114 00:09:33.700 ⇒ 00:09:36.389 Samuel Roberts: Errors or high latency, something triggers.

115 00:09:36.550 ⇒ 00:09:37.310 Mustafa Raja: Oh, yeah.

116 00:09:37.310 ⇒ 00:09:46.030 Samuel Roberts: So I think that’s probably part of this eval’s logging, where… If something goes…

117 00:09:46.610 ⇒ 00:09:51.450 Samuel Roberts: Wrong… like, just making sure that there’s, like, a webhook that we can call or something.

118 00:09:51.650 ⇒ 00:09:54.619 Samuel Roberts: Or, you know, whatever it is, like, however we’re tapping into that.

119 00:09:55.780 ⇒ 00:10:01.200 Samuel Roberts: So I don’t know if that happens in Maestra, I don’t know if that happens in… the…

120 00:10:01.460 ⇒ 00:10:05.559 Samuel Roberts: BigQuery, but I bet we probably want it closer to the source than anything.

121 00:10:06.870 ⇒ 00:10:09.410 Mustafa Raja: Yeah, then it will be in Masra, right?

122 00:10:10.020 ⇒ 00:10:11.849 Samuel Roberts: Yeah, I think that’s probably the best bet.

123 00:10:11.850 ⇒ 00:10:19.520 Amber Lin: Okay, so that would be… is that also gonna be… are we gonna do a spike this week to figure out where it is, or are we pushing it back?

124 00:10:23.870 ⇒ 00:10:28.829 Samuel Roberts: I think we can probably spike in this week, Mustafa, do you think you’ll be able to just, like, take a look and see?

125 00:10:28.830 ⇒ 00:10:37.419 Mustafa Raja: Yeah, I think Spike would be nice, because I’m thinking right now, like, where in the codebase is the best way to hook this up? So I think a spike would be nice.

126 00:10:39.090 ⇒ 00:10:46.099 Amber Lin: Cool, okay. I don’t think this… we need to grab the ticket there. Let’s see…

127 00:10:54.590 ⇒ 00:11:02.680 Amber Lin: Sorry, this is… Evil lobbying… Is this?

128 00:11:02.800 ⇒ 00:11:05.329 Amber Lin: Okay, well, we’ll update it today.

129 00:11:05.330 ⇒ 00:11:07.630 Samuel Roberts: Yeah, you’ve got updates when you’re, yeah.

130 00:11:07.630 ⇒ 00:11:17.740 Amber Lin: Okay, alright, let’s see… What about these two that’s also, like, in… In progress.

131 00:11:21.120 ⇒ 00:11:25.479 Samuel Roberts: I mean, load testing probably has to happen once it’s up in Google, so we probably have to bump that.

132 00:11:25.480 ⇒ 00:11:28.490 Amber Lin: Okay, so that will be bumped after this.

133 00:11:29.540 ⇒ 00:11:30.180 Samuel Roberts: Right?

134 00:11:30.180 ⇒ 00:11:35.279 Amber Lin: After… I think, or after this?

135 00:11:44.440 ⇒ 00:11:45.810 Samuel Roberts: Yeah, that’s probably good.

136 00:11:46.260 ⇒ 00:11:53.990 Amber Lin: Okay, so we’ll leave it there. What about, like, performance and correctness issues? Does that also go later?

137 00:11:53.990 ⇒ 00:11:58.619 Samuel Roberts: Yeah, that’s kind of like an ongoing, like, as we see things, I don’t think there’s… you know, that’s kind of dependent on that.

138 00:11:59.330 ⇒ 00:12:04.049 Samuel Roberts: But it’s not, yeah, it’s less of a, like, immediate task, and, like, tickets will come out of that from… yeah.

139 00:12:04.050 ⇒ 00:12:04.580 Amber Lin: Okay.

140 00:12:04.730 ⇒ 00:12:18.120 Amber Lin: Cool. So, okay, that leaves us with a lighter load. I think we’ll look more on track this way. Let’s just do a quick check of the next few weeks, if that… if we have enough capacity.

141 00:12:18.120 ⇒ 00:12:28.220 Amber Lin: Assuming that we close these out this week, so next week nothing really runs over, so we’ll have, like… this is also, like, an ongoing thing.

142 00:12:28.220 ⇒ 00:12:29.040 Samuel Roberts: Yeah, exactly.

143 00:12:29.040 ⇒ 00:12:35.280 Amber Lin: And then we’ll have master logs, dbt, which is less… Go ahead.

144 00:12:35.280 ⇒ 00:12:39.440 Mustafa Raja: So, for master logs, I’m seeing, if you scroll down a little.

145 00:12:39.920 ⇒ 00:12:49.090 Mustafa Raja: to 47. Yeah, 47 says, set up Cloud SQL. This needs to happen before that.

146 00:12:51.270 ⇒ 00:12:57.199 Amber Lin: Mmm, we have them in parallel, so, like, this needs to go…

147 00:12:57.560 ⇒ 00:13:00.949 Amber Lin: But we can’t… we can’t start that this week, right? It has to happen next.

148 00:13:00.950 ⇒ 00:13:06.140 Mustafa Raja: Yeah, yeah, yeah, and I’m also, hmm… I’m also wondering…

149 00:13:07.070 ⇒ 00:13:14.859 Mustafa Raja: it doesn’t make sense for us to, push to master logs until we, you know, we are done with the cloud run, right?

150 00:13:17.380 ⇒ 00:13:21.679 Amber Lin: How would that affect our, like, experimentation with, like, or testing?

151 00:13:22.020 ⇒ 00:13:23.759 Amber Lin: If we don’t have the logs.

152 00:13:26.110 ⇒ 00:13:26.510 Mustafa Raja: We don’t.

153 00:13:26.510 ⇒ 00:13:29.279 Samuel Roberts: Yeah, we’re still gonna want logs, but it might just be production logs.

154 00:13:29.690 ⇒ 00:13:30.270 Amber Lin: Okay.

155 00:13:30.270 ⇒ 00:13:44.040 Mustafa Raja: What I’m trying to say is we can only… I believe we can only push to, push logs to, BigQuery once we have, anti-deployed to Cloud Run, you know?

156 00:13:44.790 ⇒ 00:13:50.379 Amber Lin: Okay, so this will have to happen after, like, this will happen, and then that will happen, right?

157 00:13:50.740 ⇒ 00:13:54.109 Mustafa Raja: Yeah, yeah. Let me know what you think about that, Sam.

158 00:13:54.110 ⇒ 00:13:57.499 Samuel Roberts: I think, I think it’s fine, we can still figure out how to do it and be ready.

159 00:14:00.020 ⇒ 00:14:08.199 Amber Lin: Okay. Sounds like, like, master logs to BQ and migrate logging to Snowflake to BQ, they’re, like, pretty similar things? Like, does the.

160 00:14:08.200 ⇒ 00:14:08.690 Mustafa Raja: dbt.

161 00:14:08.690 ⇒ 00:14:11.910 Amber Lin: need to happen, like, right here, or can that be…

162 00:14:12.070 ⇒ 00:14:14.969 Amber Lin: Is that less important and can happen later?

163 00:14:15.370 ⇒ 00:14:18.659 Amber Lin: Cause this is gonna be a busy week, sounds like.

164 00:14:18.660 ⇒ 00:14:19.050 Samuel Roberts: Yeah.

165 00:14:19.050 ⇒ 00:14:26.069 Mustafa Raja: Yeah, so we already have the BigQuery instance, so I think dbt can stay there.

166 00:14:28.390 ⇒ 00:14:29.410 Samuel Roberts: Okay, so we’re…

167 00:14:29.410 ⇒ 00:14:31.029 Amber Lin: Yeah, because migrating is different.

168 00:14:31.440 ⇒ 00:14:32.079 Mustafa Raja: Yeah, yeah.

169 00:14:32.080 ⇒ 00:14:35.189 Samuel Roberts: Yeah, migrating is a different, like, path here.

170 00:14:35.190 ⇒ 00:14:40.230 Mustafa Raja: By migrating, do we mean that we are going to grab all of the stuff?

171 00:14:40.360 ⇒ 00:14:43.650 Mustafa Raja: That’s in Snowflake, and push that to BigQuery?

172 00:14:44.660 ⇒ 00:14:45.690 Mustafa Raja: Because I was thinking.

173 00:14:45.690 ⇒ 00:14:48.080 Samuel Roberts: Oh, all the historical data? Yeah, I know.

174 00:14:48.080 ⇒ 00:14:50.109 Mustafa Raja: I need to change the schema also.

175 00:14:50.760 ⇒ 00:14:55.180 Samuel Roberts: Yeah, we probably want to just move all the, like, consider that the N8N

176 00:14:56.170 ⇒ 00:14:58.979 Samuel Roberts: logs. I don’t think we need to move everything, do we?

177 00:14:59.240 ⇒ 00:15:00.079 Mustafa Raja: Yeah, I don’t think so either.

178 00:15:00.080 ⇒ 00:15:02.059 Samuel Roberts: It’s not yet. That’s not as pressing.

179 00:15:02.200 ⇒ 00:15:02.990 Mustafa Raja: Yeah, okay.

180 00:15:04.490 ⇒ 00:15:09.430 Amber Lin: Cool. Okay, so this week we will… bomb.

181 00:15:09.980 ⇒ 00:15:13.399 Amber Lin: I think the Texas sequel needs to go, like, another week, right?

182 00:15:13.880 ⇒ 00:15:16.140 Amber Lin: I don’t think we can get it done this week.

183 00:15:16.480 ⇒ 00:15:17.120 Samuel Roberts: No.

184 00:15:17.500 ⇒ 00:15:23.020 Amber Lin: Okay, so then, next week we have… looks like we will initialize

185 00:15:23.130 ⇒ 00:15:27.360 Amber Lin: Dbt… are we wrapping up this one this week?

186 00:15:28.990 ⇒ 00:15:32.670 Casie Aviles: That should be my focus for this week.

187 00:15:32.850 ⇒ 00:15:33.650 Amber Lin: Cool, okay.

188 00:15:33.960 ⇒ 00:15:39.490 Amber Lin: So I will extend that. So we’ll wrap that up. Next week, we will have,

189 00:15:40.070 ⇒ 00:15:56.410 Amber Lin: initialized DPT, and then we’ll have, like, text-to-SQL, which is our high priority stuff, and then, like, oh, by the way, feedback, capture, and logging, and then, like, this is also, I think, our top priority, because it sounds like it’s blocking some other stuff.

190 00:15:56.980 ⇒ 00:16:01.189 Mustafa Raja: Yeah, so, could you scroll… can you scroll up a little?

191 00:16:01.190 ⇒ 00:16:01.720 Amber Lin: Yeah.

192 00:16:03.030 ⇒ 00:16:10.380 Mustafa Raja: a little more… Okay, so, this 10 really seemed… this 10…

193 00:16:11.240 ⇒ 00:16:19.189 Mustafa Raja: And the one that we see bef- the one… the one that we put in high priority, I think these are the same thing, no?

194 00:16:19.520 ⇒ 00:16:20.060 Amber Lin: Which one?

195 00:16:20.060 ⇒ 00:16:22.790 Mustafa Raja: They’re related, certainly, yeah. So, 10…

196 00:16:23.450 ⇒ 00:16:24.210 Amber Lin: Pen, and…

197 00:16:24.210 ⇒ 00:16:24.850 Samuel Roberts: And…

198 00:16:25.050 ⇒ 00:16:26.560 Mustafa Raja: And then scroll down.

199 00:16:26.970 ⇒ 00:16:28.230 Mustafa Raja: The Falcon.

200 00:16:29.480 ⇒ 00:16:38.780 Amber Lin: 46. How would we name them differently? Like, I… like, I would love to consolidate that, I just don’t know how to name them.

201 00:16:42.280 ⇒ 00:16:45.990 Amber Lin: Because that sounds like Mansur, and that sounds like Cloud. Are they.

202 00:16:45.990 ⇒ 00:16:49.990 Samuel Roberts: Well, the master’s running on the cloud, yeah, it will be running on the cloud run, yeah.

203 00:16:52.440 ⇒ 00:16:53.900 Samuel Roberts: Okay, so…

204 00:16:54.170 ⇒ 00:16:55.759 Amber Lin: Dom, by the way, I guess.

205 00:16:55.760 ⇒ 00:16:58.499 Samuel Roberts: Yeah, yeah, we’re… yeah, that milestone’s fucked up.

206 00:16:58.770 ⇒ 00:16:59.350 Amber Lin: Okay.

207 00:17:02.370 ⇒ 00:17:04.359 Samuel Roberts: Okay, what about,

208 00:17:05.440 ⇒ 00:17:11.260 Samuel Roberts: I mean, Casey, how are you thinking you implement Monster Endpoint versus the Cloud Run? Like, they seem very…

209 00:17:12.319 ⇒ 00:17:13.579 Samuel Roberts: Yay!

210 00:17:13.749 ⇒ 00:17:15.289 Casie Aviles: Yeah, they do some…

211 00:17:15.659 ⇒ 00:17:22.099 Casie Aviles: Because, essentially, what I’m trying to do is, I’m trying to deploy our master app on Google Cloud Run.

212 00:17:22.520 ⇒ 00:17:22.920 Amber Lin: Mmm…

213 00:17:24.079 ⇒ 00:17:27.339 Samuel Roberts: No. So they’re kind of the same thing, there might be a little bit of, like.

214 00:17:28.010 ⇒ 00:17:35.440 Amber Lin: So, we can actually… I guess I’ll expand this ticket and just say, like… deploy a manager on cloud.

215 00:17:35.900 ⇒ 00:17:40.029 Amber Lin: Like, not an entry endpoint, just, like, entry endpoint of cloud.

216 00:17:40.290 ⇒ 00:17:41.670 Amber Lin: I’ll put over here.

217 00:17:41.910 ⇒ 00:17:43.160 Casie Aviles: Does that work?

218 00:17:44.190 ⇒ 00:17:51.029 Amber Lin: Okay So… I’m gonna move like this.

219 00:17:51.770 ⇒ 00:17:53.500 Amber Lin: over… here.

220 00:17:55.330 ⇒ 00:17:56.120 Amber Lin: Is that okay?

221 00:17:56.520 ⇒ 00:17:57.670 Mustafa Raja: Hmm, yeah.

222 00:17:58.890 ⇒ 00:18:00.250 Mustafa Raja: Service accounts.

223 00:18:00.910 ⇒ 00:18:05.389 Amber Lin: And then do that, and then finish up, like, what’s the difference now between.

224 00:18:05.390 ⇒ 00:18:08.709 Samuel Roberts: Do we need that configured Cloud Run, or can we just merge that into…

225 00:18:08.710 ⇒ 00:18:10.449 Mustafa Raja: Yeah, I think… Happy to.

226 00:18:12.940 ⇒ 00:18:15.230 Amber Lin: You said it blocks these. Is this still blocked?

227 00:18:15.230 ⇒ 00:18:20.470 Samuel Roberts: Yeah, no, I would say keep, keep implementing Moscra and get rid of, 46.

228 00:18:20.650 ⇒ 00:18:23.330 Samuel Roberts: Yeah, okay. And those are effectively the same thing.

229 00:18:23.730 ⇒ 00:18:24.719 Amber Lin: Cool, great.

230 00:18:25.380 ⇒ 00:18:26.670 Amber Lin: I’ll delete this.

231 00:18:27.460 ⇒ 00:18:28.040 Samuel Roberts: Yeah.

232 00:18:29.300 ⇒ 00:18:30.900 Samuel Roberts: That’s better, that’s better.

233 00:18:31.870 ⇒ 00:18:33.060 Amber Lin: Alright, so we have that.

234 00:18:33.060 ⇒ 00:18:34.989 Samuel Roberts: Does that work for you, Casey? Is that clear enough?

235 00:18:35.720 ⇒ 00:18:37.880 Casie Aviles: Yeah, I think that’s clear now.

236 00:18:41.670 ⇒ 00:18:44.220 Amber Lin: Could I move this back up to the first section?

237 00:18:44.900 ⇒ 00:18:47.360 Amber Lin: Is it easy to read this way?

238 00:18:47.640 ⇒ 00:18:49.779 Casie Aviles: I think you can keep it there.

239 00:18:49.980 ⇒ 00:18:50.950 Samuel Roberts: Money.

240 00:18:51.000 ⇒ 00:18:53.859 Casie Aviles: Because it’s related to GCP.

241 00:18:54.220 ⇒ 00:18:54.890 Amber Lin: Okay.

242 00:18:55.050 ⇒ 00:18:58.580 Casie Aviles: I think we just need to call it. How can I name it better?

243 00:18:58.980 ⇒ 00:19:00.889 Amber Lin: and train in GCP.

244 00:19:01.050 ⇒ 00:19:09.039 Mustafa Raja: I think, deploy a master Andy to… Master Andy, on Cloud Run would be a nice one.

245 00:19:09.040 ⇒ 00:19:09.489 Casie Aviles: Yeah, yeah.

246 00:19:09.490 ⇒ 00:19:09.860 Samuel Roberts: That’s.

247 00:19:09.860 ⇒ 00:19:10.380 Casie Aviles: Right.

248 00:19:14.790 ⇒ 00:19:15.480 Amber Lin: Okay.

249 00:19:16.950 ⇒ 00:19:19.300 Amber Lin: Great. I think that’s more clear.

250 00:19:19.960 ⇒ 00:19:20.490 Samuel Roberts: Yeah.

251 00:19:20.680 ⇒ 00:19:22.440 Samuel Roberts: Instead of getting to it.

252 00:19:22.620 ⇒ 00:19:25.829 Amber Lin: Okay, so we’ll have that, and then… yes, go ahead.

253 00:19:25.830 ⇒ 00:19:29.380 Casie Aviles: scroll a bit up, I think we might need to pull something.

254 00:19:30.100 ⇒ 00:19:33.290 Casie Aviles: Yeah, the edge case, right? Should we pull this down as well?

255 00:19:33.980 ⇒ 00:19:35.130 Amber Lin: Yeah, let me do that.

256 00:19:36.180 ⇒ 00:19:38.969 Amber Lin: So, that will also go in here.

257 00:19:39.470 ⇒ 00:19:44.700 Amber Lin: Does that… does the edge case… Okay, let’s see…

258 00:19:45.170 ⇒ 00:19:52.079 Amber Lin: webhook point, staging access, da-da-da, so this would, like, stay here, right? It was still…

259 00:19:52.280 ⇒ 00:19:54.110 Amber Lin: Go after we implement it?

260 00:19:54.110 ⇒ 00:19:55.780 Samuel Roberts: The… yeah.

261 00:19:57.110 ⇒ 00:19:58.080 Samuel Roberts: That’s fine.

262 00:19:58.640 ⇒ 00:20:03.930 Amber Lin: Cool. And then this will, I guess, will keep running until we… I go there.

263 00:20:04.940 ⇒ 00:20:05.890 Amber Lin: Okay.

264 00:20:06.560 ⇒ 00:20:15.400 Amber Lin: Is this… does this need to be a separate, like… Section?

265 00:20:15.830 ⇒ 00:20:16.949 Amber Lin: Or is this something…

266 00:20:16.950 ⇒ 00:20:18.580 Mustafa Raja: Related somewhere.

267 00:20:18.580 ⇒ 00:20:23.100 Samuel Roberts: Yeah, that’s something that’s gonna be… it’s gonna get, split apart more eventually.

268 00:20:23.100 ⇒ 00:20:23.660 Amber Lin: Oh, okay.

269 00:20:23.660 ⇒ 00:20:26.919 Samuel Roberts: We get closer to it. Cool, yeah. It’s a whole project in and of itself.

270 00:20:26.920 ⇒ 00:20:27.380 Mustafa Raja: Yeah.

271 00:20:27.380 ⇒ 00:20:31.159 Amber Lin: I see, and this is gonna happen, like, right after…

272 00:20:31.160 ⇒ 00:20:33.119 Mustafa Raja: Yeah. These, or is it gonna…

273 00:20:34.140 ⇒ 00:20:38.339 Mustafa Raja: Yeah, this has to happen after that, we need to have the database there.

274 00:20:38.820 ⇒ 00:20:40.060 Amber Lin: Okay, cool.

275 00:20:40.470 ⇒ 00:20:43.710 Amber Lin: So, this week we’ll finish up these, I think.

276 00:20:43.850 ⇒ 00:20:48.209 Amber Lin: I think this is, like, a good workload, right? We only have, like, 1, 2, 3, 4.

277 00:20:48.730 ⇒ 00:20:49.900 Amber Lin: Poor things.

278 00:20:49.900 ⇒ 00:20:52.200 Samuel Roberts: Yeah, they’re pretty, pretty big ones, too.

279 00:20:52.200 ⇒ 00:21:03.089 Amber Lin: things, yeah. And then next week, we have dbt, we have these, like, slightly smaller things, and then,

280 00:21:03.280 ⇒ 00:21:07.999 Amber Lin: like, I think this is a big task. Once we get that done, like, we’ll start on the…

281 00:21:08.410 ⇒ 00:21:13.439 Amber Lin: like, this is… I think next week, like, that will be the core thing, like, the cloud SQL.

282 00:21:13.560 ⇒ 00:21:14.480 Amber Lin: Staff.

283 00:21:14.670 ⇒ 00:21:30.250 Mustafa Raja: Yeah, I have a question regarding this. So, I’ve just implemented, local Postgres, right, and Master isn’t going to build with that, unless we have a, you know, valid…

284 00:21:31.020 ⇒ 00:21:34.699 Mustafa Raja: Postgres on the server, so… Sure.

285 00:21:35.210 ⇒ 00:21:41.619 Mustafa Raja: Would you reckon that I, push it back to, SQLite?

286 00:21:42.340 ⇒ 00:21:49.030 Mustafa Raja: Or… should we bump this up, the SQL? Should we do it this week?

287 00:21:49.680 ⇒ 00:21:51.560 Samuel Roberts: Think…

288 00:21:53.030 ⇒ 00:22:00.839 Samuel Roberts: I think for now, since we’re just, like, this isn’t going live production, we can probably even just use, like, Supabase or something if you need a cloud one.

289 00:22:01.080 ⇒ 00:22:01.430 Mustafa Raja: Yeah.

290 00:22:01.570 ⇒ 00:22:04.709 Samuel Roberts: And then once we get the other one set up, we can move it over.

291 00:22:04.900 ⇒ 00:22:05.530 Mustafa Raja: Okay.

292 00:22:05.910 ⇒ 00:22:18.390 Mustafa Raja: Yeah, then I… Does that make sense? Does that work? Yeah, that does make sense. I will then connect, connect our local Maestra that we have, to Superbase, ABC project then, right?

293 00:22:18.880 ⇒ 00:22:21.010 Samuel Roberts: Yeah, just like a new… is it one table?

294 00:22:21.530 ⇒ 00:22:27.439 Mustafa Raja: It’s… let me see how many… it’s, 1, 2, 3, 4, 7… Oh, eat, actually.

295 00:22:27.440 ⇒ 00:22:30.369 Samuel Roberts: 8 more tables, 8 more tables? I’m gonna say that.

296 00:22:30.830 ⇒ 00:22:31.859 Mustafa Raja: Which heaven, yeah.

297 00:22:32.120 ⇒ 00:22:33.840 Mustafa Raja: 1, 2, 3… yeah, 7.

298 00:22:37.800 ⇒ 00:22:43.149 Samuel Roberts: I’m trying to think if that’s even worth it, or if just the SQLite is fine, and then we’ll flip back to that later.

299 00:22:47.200 ⇒ 00:22:52.899 Samuel Roberts: So, the issue with the SQLite is that it won’t stay between refreshes, or, restarts, right?

300 00:22:52.900 ⇒ 00:22:53.530 Mustafa Raja: Yeah.

301 00:22:54.430 ⇒ 00:22:59.169 Samuel Roberts: Yeah, let’s say for… for dev, that might be okay, maybe we just sync it down or something?

302 00:23:00.580 ⇒ 00:23:11.709 Samuel Roberts: I don’t know if it’s worth making 8 more tables in Supabase for that, if we’re gonna move it eventually. Yeah. Okay, yeah, I would say I’m glad we have that ready to go, but let’s, yeah, flip back to the SQLite for now.

303 00:23:12.200 ⇒ 00:23:12.860 Mustafa Raja: Yeah.

304 00:23:13.460 ⇒ 00:23:14.520 Mustafa Raja: Okay.

305 00:23:14.830 ⇒ 00:23:15.719 Mustafa Raja: I’ll do that.

306 00:23:16.630 ⇒ 00:23:17.120 Samuel Roberts: Okay. Cool.

307 00:23:17.810 ⇒ 00:23:20.450 Amber Lin: Will this also take 2 weeks?

308 00:23:21.740 ⇒ 00:23:23.930 Mustafa Raja: Oh, I don’t think so, to be honest.

309 00:23:24.610 ⇒ 00:23:25.990 Amber Lin: So…

310 00:23:25.990 ⇒ 00:23:27.109 Samuel Roberts: We can probably bump that back a week.

311 00:23:27.110 ⇒ 00:23:32.670 Amber Lin: Can… can we clean it? Like, are we… do you think we’re confident to close it out in one week?

312 00:23:32.670 ⇒ 00:23:33.300 Mustafa Raja: Yep.

313 00:23:33.680 ⇒ 00:23:35.559 Amber Lin: Cool, okay.

314 00:23:35.560 ⇒ 00:23:38.330 Casie Aviles: Might need to spike a bit on that.

315 00:23:38.330 ⇒ 00:23:40.079 Mustafa Raja: As well, since…

316 00:23:40.850 ⇒ 00:23:44.410 Samuel Roberts: Yeah, maybe back that up into this week a little bit, so we can start the spike.

317 00:23:44.690 ⇒ 00:23:45.360 Amber Lin: Okay.

318 00:23:46.220 ⇒ 00:23:49.500 Casie Aviles: Because we… we haven’t really set up…

319 00:23:49.500 ⇒ 00:23:50.180 Samuel Roberts: Yeah.

320 00:23:50.180 ⇒ 00:23:55.090 Casie Aviles: Postgres database in Google, so… Might need to figure that out.

321 00:23:55.550 ⇒ 00:24:03.439 Amber Lin: Glad we talked about that. Okay. And then, like, after that, we’ll have, let’s see…

322 00:24:04.160 ⇒ 00:24:09.400 Amber Lin: Master logs would be queued, so depending… dependent on, like, Cloud Run.

323 00:24:09.570 ⇒ 00:24:13.920 Amber Lin: Bq, migration, logging…

324 00:24:14.040 ⇒ 00:24:18.919 Amber Lin: Testing, thumbs up, thumbs down. Like, I feel like that… that week has a lot of stuff.

325 00:24:19.170 ⇒ 00:24:21.470 Amber Lin: Is that a reasonable workload?

326 00:24:25.910 ⇒ 00:24:27.280 Casie Aviles: This week. Let’s see…

327 00:24:34.400 ⇒ 00:24:38.400 Mustafa Raja: When does… 48 mean?

328 00:24:40.460 ⇒ 00:24:41.330 Casie Aviles: Yay.

329 00:24:43.250 ⇒ 00:24:49.759 Samuel Roberts: I think that’s just getting the, like, developer, or the dev…

330 00:24:50.350 ⇒ 00:24:54.580 Samuel Roberts: Google Chat working and making sure everything’s good there.

331 00:24:54.580 ⇒ 00:24:55.280 Mustafa Raja: Okay.

332 00:24:55.280 ⇒ 00:24:57.309 Samuel Roberts: Then we have staging validation.

333 00:24:57.710 ⇒ 00:24:59.989 Samuel Roberts: And then we have live.

334 00:25:03.380 ⇒ 00:25:06.779 Samuel Roberts: But I think Dev is, like, just our… our chat, and then…

335 00:25:07.150 ⇒ 00:25:10.100 Samuel Roberts: Staging might be their staging chat, right?

336 00:25:10.480 ⇒ 00:25:11.000 Mustafa Raja: Yeah, yeah.

337 00:25:11.000 ⇒ 00:25:11.930 Amber Lin: Makes sense.

338 00:25:12.680 ⇒ 00:25:17.359 Amber Lin: So… Sounds like these are more, like.

339 00:25:17.840 ⇒ 00:25:24.409 Amber Lin: Is this doable in one week? Like, 1, 2, 3… 4, 5, 6, 7.

340 00:25:26.080 ⇒ 00:25:31.999 Amber Lin: Does this go here or go later? I feel like this needs to, like, thumbs up and down need to have it before.

341 00:25:32.000 ⇒ 00:25:33.190 Samuel Roberts: Yeah, that’s fine.

342 00:25:33.190 ⇒ 00:25:33.820 Amber Lin: Validate.

343 00:25:33.820 ⇒ 00:25:39.840 Samuel Roberts: deploy master service to client GCP is kind of also a thing that’s… Yeah, before.

344 00:25:40.110 ⇒ 00:25:40.680 Mustafa Raja: Yeah.

345 00:25:40.980 ⇒ 00:25:46.329 Samuel Roberts: Yeah, I think that’s a duplicate. Like, it might be dev staging prob, might be a little different, but I don’t think we need a whole…

346 00:25:46.560 ⇒ 00:25:47.270 Samuel Roberts: Thing for that.

347 00:25:47.270 ⇒ 00:25:52.549 Amber Lin: So, like, that would be probably, like, this one, this one, this one, so I can… we’ll just delete it.

348 00:25:52.550 ⇒ 00:25:54.470 Samuel Roberts: Yeah, just delete that one.

349 00:25:54.780 ⇒ 00:25:55.650 Amber Lin: Okay, great.

350 00:26:00.940 ⇒ 00:26:02.130 Amber Lin: Can…

351 00:26:02.130 ⇒ 00:26:02.890 Casie Aviles: It’s be a…

352 00:26:02.890 ⇒ 00:26:10.509 Amber Lin: Same ticket, like, dev, development, and then validate? Are they… do they have to be separate?

353 00:26:13.060 ⇒ 00:26:20.099 Samuel Roberts: Which one? Establish, Andy? Establish and validate. Validate. Yeah, you can make it establish and validate, that’s fine.

354 00:26:20.400 ⇒ 00:26:21.260 Casie Aviles: illness.

355 00:26:22.890 ⇒ 00:26:24.110 Amber Lin: Okay, cool.

356 00:26:30.600 ⇒ 00:26:37.870 Amber Lin: Cool. I think that’s more straightforward of, like, staging, dev, and then we have…

357 00:26:38.840 ⇒ 00:26:42.630 Amber Lin: Answer logs to BQ, from Snowflake to BQ.

358 00:26:43.230 ⇒ 00:26:45.879 Amber Lin: What’s the difference between these?

359 00:26:47.150 ⇒ 00:26:48.429 Amber Lin: this, this…

360 00:26:49.680 ⇒ 00:26:50.970 Casie Aviles: Different is.

361 00:26:51.660 ⇒ 00:26:53.579 Casie Aviles: 15 at 16.

362 00:26:54.010 ⇒ 00:26:57.209 Amber Lin: Yeah, library logging probably should be cute.

363 00:26:59.510 ⇒ 00:27:02.130 Samuel Roberts: Those might not be different now if we’re not moving all the data.

364 00:27:04.890 ⇒ 00:27:06.830 Samuel Roberts: There might be the… there might be duplicates.

365 00:27:06.830 ⇒ 00:27:07.480 Mustafa Raja: Yeah.

366 00:27:08.260 ⇒ 00:27:09.480 Amber Lin: Okay, so…

367 00:27:09.480 ⇒ 00:27:13.769 Samuel Roberts: I think there was kind of a question before about whether or not, like, how the logging was gonna go, but since…

368 00:27:15.060 ⇒ 00:27:16.149 Samuel Roberts: I think, yeah, you can probably.

369 00:27:16.150 ⇒ 00:27:20.990 Amber Lin: Okay, I’m gonna delete that one. Is this the best way to name it right now?

370 00:27:23.580 ⇒ 00:27:25.430 Amber Lin: Are we migrating logging?

371 00:27:26.620 ⇒ 00:27:30.109 Samuel Roberts: I would say it would just be master logs to BQ or something.

372 00:27:35.470 ⇒ 00:27:38.250 Amber Lin: Am I spelling Mastra correctly?

373 00:27:38.250 ⇒ 00:27:39.120 Casie Aviles: There’s no end.

374 00:27:39.120 ⇒ 00:27:42.220 Samuel Roberts: No answer. Okay, that’s good.

375 00:27:42.220 ⇒ 00:27:44.909 Amber Lin: Cool, okay, that… that’s nice, so…

376 00:27:45.100 ⇒ 00:27:49.779 Amber Lin: 2BQ, established, staging, established development.

377 00:27:50.030 ⇒ 00:27:54.720 Amber Lin: testing. I think that’s… that’s, like, that looks reasonable. Do you guys think so?

378 00:27:56.350 ⇒ 00:27:56.940 Mustafa Raja: Yep.

379 00:27:57.410 ⇒ 00:27:58.539 Amber Lin: Or should we move this?

380 00:27:59.060 ⇒ 00:28:01.860 Samuel Roberts: No, I think that’s good. I think we’re gonna want that. Okay.

381 00:28:02.220 ⇒ 00:28:03.350 Amber Lin: Okay, cool.

382 00:28:03.520 ⇒ 00:28:07.269 Amber Lin: And then, like, afterwards, I think, I think that’s reasonable.

383 00:28:08.620 ⇒ 00:28:11.250 Amber Lin: Actually, I don’t know if that’s reasonable.

384 00:28:11.250 ⇒ 00:28:13.700 Samuel Roberts: Nope, go back over for a second.

385 00:28:13.890 ⇒ 00:28:18.040 Samuel Roberts: This end-to-end depreciation, is that the right… that doesn’t seem like a good date for that.

386 00:28:19.220 ⇒ 00:28:22.140 Amber Lin: And deprecation… which one?

387 00:28:22.920 ⇒ 00:28:24.370 Samuel Roberts: Up here, 40.

388 00:28:25.590 ⇒ 00:28:27.730 Amber Lin: for the… Oh.

389 00:28:27.730 ⇒ 00:28:28.360 Mustafa Raja: Oh, no.

390 00:28:28.360 ⇒ 00:28:31.390 Amber Lin: Does that need to go, like… After this… Probably back then.

391 00:28:31.390 ⇒ 00:28:32.980 Samuel Roberts: Yeah, I think so.

392 00:28:32.980 ⇒ 00:28:37.140 Amber Lin: This would need to go… like, here?

393 00:28:37.140 ⇒ 00:28:40.400 Samuel Roberts: Because we’re not gonna turn that off until we’re ready to turn the other one on, so…

394 00:28:40.710 ⇒ 00:28:44.869 Amber Lin: Okay, so that will happen after we provide staging access.

395 00:28:45.660 ⇒ 00:28:46.150 Samuel Roberts: Yeah, probably.

396 00:28:46.150 ⇒ 00:28:46.750 Mustafa Raja: Yeah.

397 00:28:46.750 ⇒ 00:28:49.399 Casie Aviles: Okay. These are the milestones, right?

398 00:28:50.280 ⇒ 00:28:51.570 Casie Aviles: Huh. Yeah.

399 00:28:51.680 ⇒ 00:28:53.410 Amber Lin: This one, and then this one.

400 00:28:53.880 ⇒ 00:28:59.039 Amber Lin: Then this shouldn’t be there, right? And then deprecation should be…

401 00:28:59.040 ⇒ 00:29:03.249 Samuel Roberts: Yeah, that should be further… yeah, exactly. Yeah, it even says retire NADN, and that… yeah, so that makes sense.

402 00:29:03.250 ⇒ 00:29:04.030 Amber Lin: Okay.

403 00:29:05.450 ⇒ 00:29:06.250 Samuel Roberts: That’s better, okay.

404 00:29:07.360 ⇒ 00:29:12.599 Amber Lin: Alright, I like that. This… this week.

405 00:29:13.550 ⇒ 00:29:19.410 Amber Lin: Real, are we sticking with real, redeem Omni, or are we just keeping real?

406 00:29:20.020 ⇒ 00:29:22.640 Casie Aviles: We haven’t really had a conversation about that.

407 00:29:22.640 ⇒ 00:29:30.700 Amber Lin: I see. Gotcha. Okay, I’ll leave it, I’ll leave it there. Love testing, rank, this is ongoing. Slack alerting, ongoing.

408 00:29:31.030 ⇒ 00:29:38.789 Amber Lin: Okay, I think that’s reasonable. Cool. Let’s clean up Linear a little bit.

409 00:29:42.100 ⇒ 00:29:44.990 Amber Lin: Okay.

410 00:29:47.180 ⇒ 00:29:52.200 Amber Lin: How should we approach these? Let’s say… Save the…

411 00:29:52.200 ⇒ 00:29:55.610 Mustafa Raja: We can move it to the 1569.

412 00:29:56.740 ⇒ 00:29:59.009 Mustafa Raja: I think you reviewed it, and it was good.

413 00:29:59.010 ⇒ 00:29:59.810 Samuel Roberts: Yeah.

414 00:30:01.470 ⇒ 00:30:09.100 Mustafa Raja: And then, yeah, this, update, this 1721. So, a quick update on that,

415 00:30:09.910 ⇒ 00:30:23.479 Mustafa Raja: the model’s, rate limits are shared, so we have, two deployments for GPT-4.0, and Azure provides, 30 million, tokens per minute.

416 00:30:23.480 ⇒ 00:30:40.439 Mustafa Raja: And that is shared between the two deployments, right? So, the zoom backfill, has 29 million, 29.8 million, and then the rest of the 200,000 is going to GPT-4.0.

417 00:30:40.440 ⇒ 00:30:50.519 Mustafa Raja: So Casey and I thought that we’ll just make, GPT-40 zoom backfill as our standard model, and,

418 00:30:50.550 ⇒ 00:30:55.150 Mustafa Raja: We’ll just remove GPT-40 from our N10 workflows.

419 00:30:55.660 ⇒ 00:30:59.120 Mustafa Raja: replace it with GPT-4 or Zoom Backfill. What do you think about that?

420 00:31:00.200 ⇒ 00:31:04.870 Samuel Roberts: I think that’s fine, yeah, I saw the messages on, or the linear comment.

421 00:31:05.710 ⇒ 00:31:11.290 Samuel Roberts: I took a look at it, I think that makes sense. My only question is, do we need that other one for anything, or can we spin that down?

422 00:31:11.970 ⇒ 00:31:20.370 Mustafa Raja: I think, once we know that we have, transferred over to Zoom Backfill, then we can, you know, remove that.

423 00:31:20.370 ⇒ 00:31:27.060 Samuel Roberts: Yep, that’s fine, I just want to make sure that, yeah, I didn’t know if anything else was on that. I had another question, and I don’t remember what it was.

424 00:31:28.930 ⇒ 00:31:35.820 Samuel Roberts: Oh, the other thing I guess this kind of brings up is the models eventually hosting them somewhere else.

425 00:31:36.250 ⇒ 00:31:37.380 Casie Aviles: Oh, yeah.

426 00:31:37.960 ⇒ 00:31:40.019 Amber Lin: Oh, do we have that on the Gantt?

427 00:31:40.230 ⇒ 00:31:41.040 Amber Lin: At all?

428 00:31:41.040 ⇒ 00:31:43.440 Samuel Roberts: No, no, I haven’t… we haven’t discussed that at all.

429 00:31:43.440 ⇒ 00:31:47.209 Amber Lin: I see. When do we want the conversation? I’ll put a placeholder somewhere.

430 00:31:49.720 ⇒ 00:31:53.249 Amber Lin: Like, here, or should I…

431 00:31:55.080 ⇒ 00:31:59.610 Casie Aviles: Because I… I guess the thing there is it’s not blocking…

432 00:31:59.880 ⇒ 00:32:00.820 Casie Aviles: I guess we’re… we can use.

433 00:32:00.820 ⇒ 00:32:01.430 Samuel Roberts: No, no.

434 00:32:02.080 ⇒ 00:32:02.620 Casie Aviles: Bye. Thank you.

435 00:32:02.620 ⇒ 00:32:08.030 Samuel Roberts: using ours, but since we’re moving everything to their GCP and everything, we probably want to move that as well.

436 00:32:08.570 ⇒ 00:32:09.590 Casie Aviles: Yeah, okay.

437 00:32:10.610 ⇒ 00:32:14.610 Amber Lin: So I’ll put it… I think I’ll put it, like, down here.

438 00:32:15.210 ⇒ 00:32:20.409 Amber Lin: I’ll say migrate… Models.

439 00:32:21.520 ⇒ 00:32:23.040 Amber Lin: AI models?

440 00:32:23.770 ⇒ 00:32:24.540 Casie Aviles: Yeah.

441 00:32:24.920 ⇒ 00:32:25.580 Amber Lin: Okay.

442 00:32:27.000 ⇒ 00:32:29.489 Amber Lin: Before NAN deprecation?

443 00:32:30.730 ⇒ 00:32:33.090 Amber Lin: Or before staging access to team.

444 00:32:36.190 ⇒ 00:32:36.830 Samuel Roberts: My mom…

445 00:32:36.830 ⇒ 00:32:38.499 Mustafa Raja: Access, looks good.

446 00:32:38.900 ⇒ 00:32:42.589 Amber Lin: Okay, so I’ll put it here so we discuss it. That will be…

447 00:32:42.590 ⇒ 00:32:43.370 Samuel Roberts: Yeah, yeah.

448 00:32:43.370 ⇒ 00:32:45.589 Amber Lin: On 2 weeks from… two weeks from now.

449 00:32:45.590 ⇒ 00:32:51.469 Samuel Roberts: That’s fine. We probably want to put a little thought into that, because I don’t know if 4.0 can be hosted on GCP anyway.

450 00:32:52.960 ⇒ 00:32:53.590 Mustafa Raja: Oh.

451 00:32:55.250 ⇒ 00:32:59.700 Samuel Roberts: Like, that’s an OpenAI thing, so it’s an Azure thing. I don’t know if it’s a Google thing.

452 00:33:00.320 ⇒ 00:33:06.770 Casie Aviles: what Tim mentioned is, I think they’re open to having open AI models, but…

453 00:33:06.970 ⇒ 00:33:10.510 Casie Aviles: Yeah, we can… we can confirm again with them, like, what the best…

454 00:33:11.150 ⇒ 00:33:12.520 Samuel Roberts: Okay. Setup would be.

455 00:33:13.120 ⇒ 00:33:16.469 Samuel Roberts: That’s fine, yeah, I just want to make sure we’re thinking that through, but that’s perfect.

456 00:33:16.470 ⇒ 00:33:23.239 Amber Lin: Okay, great. I’m glad we talked about it. Is this done, by the way? Is this muscle done to confirm new central structure of a client?

457 00:33:23.800 ⇒ 00:33:28.450 Samuel Roberts: That, I’m not even sure what that was there for, but that’s kind of what you were talking about today.

458 00:33:29.460 ⇒ 00:33:32.370 Amber Lin: Oh, I was like, do we have everything there? Okay, so this is.

459 00:33:32.370 ⇒ 00:33:33.480 Samuel Roberts: Yeah, so I think…

460 00:33:34.680 ⇒ 00:33:37.520 Amber Lin: Cool. I’m gonna… I’m gonna move it there.

461 00:33:38.030 ⇒ 00:33:38.560 Samuel Roberts: Definitely.

462 00:33:38.560 ⇒ 00:33:40.089 Amber Lin: Okay. Cool.

463 00:33:40.190 ⇒ 00:33:44.929 Amber Lin: Let’s see, is this done? This Prinav’s ticket?

464 00:33:47.940 ⇒ 00:33:49.580 Casie Aviles: He, he, he did it.

465 00:33:49.970 ⇒ 00:33:52.150 Casie Aviles: He already provided,

466 00:33:52.630 ⇒ 00:33:55.120 Amber Lin: Cool. What do you call this? A Notion doc.

467 00:33:55.720 ⇒ 00:33:57.010 Amber Lin: I see. Okay.

468 00:33:57.200 ⇒ 00:34:03.900 Amber Lin: Performance, let’s see, what do we need?

469 00:34:05.800 ⇒ 00:34:06.770 Amber Lin: No.

470 00:34:08.860 ⇒ 00:34:09.679 Amber Lin: Okay.

471 00:34:10.620 ⇒ 00:34:16.939 Amber Lin: I think that’s not until… that’s later, right? This one’s later.

472 00:34:16.949 ⇒ 00:34:17.659 Mustafa Raja: Yep.

473 00:34:17.659 ⇒ 00:34:20.899 Amber Lin: Load testing, so I’m gonna move it back.

474 00:34:21.659 ⇒ 00:34:27.679 Amber Lin: Move it back… who’s doing… or is this also later?

475 00:34:27.809 ⇒ 00:34:33.209 Amber Lin: performance and correctness issues. This is, like, an ongoing thing, right?

476 00:34:33.600 ⇒ 00:34:40.150 Amber Lin: So I’m also gonna move that… I think I’m gonna put it… Move it back, probably.

477 00:34:40.449 ⇒ 00:34:41.049 Samuel Roberts: Yes.

478 00:34:41.050 ⇒ 00:34:44.630 Amber Lin: Okay…

479 00:34:44.830 ⇒ 00:34:51.119 Amber Lin: We wanted a spike on that. I think we also wanted a spike on… what did we want a spike on?

480 00:34:51.370 ⇒ 00:34:52.570 Amber Lin: cloud…

481 00:34:52.570 ⇒ 00:34:54.020 Samuel Roberts: SQL.

482 00:34:54.020 ⇒ 00:34:54.820 Amber Lin: Okay.

483 00:34:54.929 ⇒ 00:34:58.069 Amber Lin: Who’s doing this bike?

484 00:35:01.130 ⇒ 00:35:02.170 Mustafa Raja: I can tick that.

485 00:35:05.050 ⇒ 00:35:07.239 Amber Lin: What are we doing, SQLite?

486 00:35:08.030 ⇒ 00:35:09.709 Mustafa Raja: No, no, no, Postgres.

487 00:35:09.990 ⇒ 00:35:10.570 Amber Lin: postcard.

488 00:35:10.570 ⇒ 00:35:10.910 Samuel Roberts: S.

489 00:35:12.580 ⇒ 00:35:15.510 Amber Lin: Postgres… Okay.

490 00:35:15.640 ⇒ 00:35:18.249 Amber Lin: How many hours do you need for the spike?

491 00:35:18.860 ⇒ 00:35:26.540 Mustafa Raja: So the main goal here would be identifying what permissions do we need and how are we going to set up, right? So…

492 00:35:27.060 ⇒ 00:35:27.490 Samuel Roberts: Yeah.

493 00:35:27.490 ⇒ 00:35:28.750 Mustafa Raja: Another tool.

494 00:35:31.810 ⇒ 00:35:33.989 Samuel Roberts: Yeah, just need to get all your ducks in a row.

495 00:35:34.620 ⇒ 00:35:35.020 Amber Lin: Cool.

496 00:35:35.020 ⇒ 00:35:37.860 Samuel Roberts: Oh, two points, I’ll do…

497 00:35:37.860 ⇒ 00:35:39.659 Amber Lin: End of this week.

498 00:35:40.160 ⇒ 00:35:45.669 Amber Lin: Great. That’s that. Implement.

499 00:35:45.860 ⇒ 00:35:46.890 Amber Lin: Yeah, sure.

500 00:35:48.410 ⇒ 00:35:50.280 Amber Lin: I’m gonna rename this.

501 00:35:50.800 ⇒ 00:35:56.119 Amber Lin: And… D2. Cloud run.

502 00:35:56.460 ⇒ 00:36:03.239 Amber Lin: Cool, and this will also be done… End of this week.

503 00:36:05.210 ⇒ 00:36:07.830 Amber Lin: And… Errors…

504 00:36:08.890 ⇒ 00:36:11.840 Casie Aviles: Oh, yeah, that’s… that’s for maintenance.

505 00:36:12.240 ⇒ 00:36:15.399 Amber Lin: Yeah, are you doing these this week, or do you have time?

506 00:36:15.600 ⇒ 00:36:26.710 Casie Aviles: I’m actually doing the maintenance and the errors right now, and then I got, like, we got a message from Yvette yesterday, and she was flagging, you know, some…

507 00:36:26.830 ⇒ 00:36:28.450 Amber Lin: More errors.

508 00:36:28.470 ⇒ 00:36:30.560 Casie Aviles: And what I told her was that…

509 00:36:30.780 ⇒ 00:36:38.780 Casie Aviles: Well, we can do… we can do within the week the short-term maintenance fixes, but there’s also, like, the ongoing migration that should help.

510 00:36:39.280 ⇒ 00:36:45.310 Casie Aviles: With that, but I did, like, give a set date for that, since I told her that it’s ongoing.

511 00:36:45.540 ⇒ 00:36:56.389 Amber Lin: I see, I see. Like, we’ll probably… we should… I’ll make a short list, and then we’ll… we’ll validate, where we think each department’s at, like, once we finish

512 00:36:56.580 ⇒ 00:37:04.799 Amber Lin: these tickets, just for this week. We’re almost there. Evals in… where’s the… sorry, where are these…

513 00:37:05.690 ⇒ 00:37:07.720 Amber Lin: These 3 tickets?

514 00:37:08.600 ⇒ 00:37:10.249 Mustafa Raja: In my thing, yeah.

515 00:37:10.250 ⇒ 00:37:14.060 Amber Lin: Oh, okay, so this is this. This is also for this week?

516 00:37:15.130 ⇒ 00:37:20.679 Amber Lin: Eval logging, this week. And then, what’s this?

517 00:37:21.120 ⇒ 00:37:25.829 Mustafa Raja: Yeah, this is done, and I’m going to share it.

518 00:37:26.420 ⇒ 00:37:27.890 Mustafa Raja: Cool.

519 00:37:28.980 ⇒ 00:37:38.829 Amber Lin: Okay, and then error latency threshold hooks. Do you want a spike on that, or are we doing that? What are we doing there? Yeah, let’s add that.

520 00:37:39.200 ⇒ 00:37:39.710 Amber Lin: Okay.

521 00:37:40.090 ⇒ 00:37:50.129 Amber Lin: Okay, so let me see… If it’s in here… See, this is done.

522 00:37:50.610 ⇒ 00:37:56.540 Amber Lin: Edge cases… Where was that?

523 00:38:01.830 ⇒ 00:38:03.319 Amber Lin: Oh, Ashley…

524 00:38:06.760 ⇒ 00:38:08.970 Amber Lin: I’ll add this to this cycle.

525 00:38:09.230 ⇒ 00:38:16.899 Amber Lin: Cool. Is this done? Did we ever do this? Like, document relevance?

526 00:38:17.520 ⇒ 00:38:18.260 Amber Lin: It’s not right.

527 00:38:19.490 ⇒ 00:38:20.580 Amber Lin: I’ll say this one.

528 00:38:21.400 ⇒ 00:38:22.119 Amber Lin: It’s this one.

529 00:38:22.120 ⇒ 00:38:26.300 Samuel Roberts: That was kind of… I think that… I think that’s also kind of the thing we were just talking about, like, getting.

530 00:38:26.300 ⇒ 00:38:26.910 Amber Lin: I see.

531 00:38:26.910 ⇒ 00:38:29.950 Samuel Roberts: It’s been paid for all the departments. Cool. Yeah.

532 00:38:30.480 ⇒ 00:38:34.710 Amber Lin: Cool, cool. I’m gonna… this…

533 00:38:35.870 ⇒ 00:38:39.170 Amber Lin: Add this to the current cycle.

534 00:38:39.370 ⇒ 00:38:43.769 Amber Lin: That will not be… oh, is that gonna also be due this week?

535 00:38:43.770 ⇒ 00:38:44.610 Mustafa Raja: Do you have enough…

536 00:38:44.610 ⇒ 00:38:45.340 Amber Lin: time.

537 00:38:45.340 ⇒ 00:38:48.659 Mustafa Raja: No, I think we’re talking about a spike, right?

538 00:38:49.320 ⇒ 00:38:55.989 Amber Lin: Okay. So then we need to stretch this out, because if we’re just doing a spike.

539 00:38:58.660 ⇒ 00:38:59.590 Amber Lin: Okay.

540 00:39:16.460 ⇒ 00:39:18.430 Amber Lin: What’s the goal of the spike?

541 00:39:19.250 ⇒ 00:39:24.019 Mustafa Raja: I’m just going to see, where in the workflow

542 00:39:24.190 ⇒ 00:39:34.739 Mustafa Raja: Can we pull this information that there’s an error or latency? So we can then, move to, move, move, move this information to Slack.

543 00:39:35.300 ⇒ 00:39:36.800 Mustafa Raja: Yeah, we’re pretty nice.

544 00:39:43.070 ⇒ 00:39:50.690 Amber Lin: Cool, I’ll say this is due this week, and I’ll move the… I’ll move this one.

545 00:39:51.330 ⇒ 00:39:53.010 Amber Lin: This one and two.

546 00:39:54.030 ⇒ 00:39:55.310 Amber Lin: Next week.

547 00:39:55.970 ⇒ 00:39:57.260 Amber Lin: Cool, okay.

548 00:39:58.520 ⇒ 00:40:07.709 Amber Lin: Take a look at your tickets, does this look doable to you with all the… all the deadlines? I’m gonna move this to a different cycle.

549 00:40:07.710 ⇒ 00:40:11.500 Casie Aviles: Yeah, that one probably won’t fit this cycle.

550 00:40:14.540 ⇒ 00:40:17.670 Mustafa Raja: Yeah, 15… 53, I think, is good to go.

551 00:40:18.210 ⇒ 00:40:19.370 Mustafa Raja: What do you think, Sam?

552 00:40:23.020 ⇒ 00:40:24.379 Samuel Roberts: Yeah, I think it’s fine.

553 00:40:24.770 ⇒ 00:40:26.780 Samuel Roberts: I think we’ve… yeah, we’ve been… we’ve done that.

554 00:40:27.260 ⇒ 00:40:27.630 Amber Lin: Okay.

555 00:40:28.070 ⇒ 00:40:38.959 Mustafa Raja: Yeah, what would be the exact, deliverables for 1572 and 1571?

556 00:40:40.860 ⇒ 00:40:42.089 Mustafa Raja: This is for Sam.

557 00:40:42.920 ⇒ 00:40:51.840 Samuel Roberts: Yeah, I think it would be, logging that stuff to… the Postgres?

558 00:40:52.490 ⇒ 00:40:58.059 Mustafa Raja: Yeah, now we are moving to… what’s it called? Moving back to… because this is done, right? I’ve already done.

559 00:40:58.060 ⇒ 00:41:01.310 Samuel Roberts: I think we want to be able to turn on logging to Postgres.

560 00:41:01.790 ⇒ 00:41:03.670 Mustafa Raja: Turn on login to Postgres.

561 00:41:03.670 ⇒ 00:41:05.239 Samuel Roberts: So, like, we wanna, we wanna have…

562 00:41:05.860 ⇒ 00:41:07.760 Samuel Roberts: You have that, you have that done, right?

563 00:41:08.060 ⇒ 00:41:10.930 Mustafa Raja: Yeah, can I share my screen? We can take a look at the tables right now.

564 00:41:11.570 ⇒ 00:41:12.160 Samuel Roberts: Sure.

565 00:41:17.620 ⇒ 00:41:20.260 Mustafa Raja: Okay, so this is the workflows.

566 00:41:22.270 ⇒ 00:41:24.409 Mustafa Raja: And we have these snapshots.

567 00:41:24.960 ⇒ 00:41:28.090 Mustafa Raja: Wait, let’s take a look at this one…

568 00:41:30.640 ⇒ 00:41:39.120 Mustafa Raja: And then in the payload… This is Hoda… Execute DB…

569 00:41:42.830 ⇒ 00:41:44.020 Mustafa Raja: Target agent…

570 00:41:50.420 ⇒ 00:41:55.060 Mustafa Raja: Yeah, I’m trying to find the key here… Wait.

571 00:42:18.870 ⇒ 00:42:31.220 Mustafa Raja: Okay, so there would be a key in, in this output. I’m going to check the code, where that went. But what that key would help us do is stitch this coder.

572 00:42:31.470 ⇒ 00:42:37.900 Mustafa Raja: Back to the workflow. So this is the relevancy score that we talked about, that we are going to replace quality score with.

573 00:42:38.460 ⇒ 00:42:39.020 Samuel Roberts: Yep.

574 00:42:39.020 ⇒ 00:42:40.630 Mustafa Raja: It’s working pretty good.

575 00:42:41.690 ⇒ 00:42:44.290 Mustafa Raja: It’s a little, what’s it called, strict?

576 00:42:44.820 ⇒ 00:42:48.620 Mustafa Raja: I don’t know if you want to, you know, make it a little lenient.

577 00:42:49.340 ⇒ 00:42:52.609 Mustafa Raja: But yeah, the overall, the soft looks pretty good.

578 00:42:54.150 ⇒ 00:42:58.419 Mustafa Raja: And this is all in, local… what’s it called? Local postpress.

579 00:42:58.940 ⇒ 00:43:04.559 Samuel Roberts: Yeah, so this is the logging that I wanted to make sure we had, but we can’t really do that until we have the database.

580 00:43:04.890 ⇒ 00:43:05.670 Mustafa Raja: Yeah, NGB.

581 00:43:05.670 ⇒ 00:43:07.160 Samuel Roberts: Right? Yeah.

582 00:43:07.160 ⇒ 00:43:07.650 Mustafa Raja: So…

583 00:43:07.830 ⇒ 00:43:13.310 Samuel Roberts: We’re good to run there. I think just for now, we gotta switch it back so that we don’t log it to nothing.

584 00:43:13.440 ⇒ 00:43:14.010 Mustafa Raja: Yeah.

585 00:43:14.210 ⇒ 00:43:23.950 Mustafa Raja: Okay, okay, then maybe we can… we can move… mark that… mark that ticket as done, and I’m… right after this meeting, move… I’ll move the database back to, SQLite.

586 00:43:24.780 ⇒ 00:43:28.339 Samuel Roberts: Okay, yeah, and then that code will be ready to go, though, for when we have…

587 00:43:28.340 ⇒ 00:43:28.830 Mustafa Raja: database.

588 00:43:28.940 ⇒ 00:43:31.670 Mustafa Raja: Because right now, it’s not going to build. So, yeah.

589 00:43:31.670 ⇒ 00:43:33.520 Samuel Roberts: Right, right. Okay, perfect.

590 00:43:39.260 ⇒ 00:43:40.390 Samuel Roberts: Alright, what else?

591 00:43:41.010 ⇒ 00:43:44.680 Mustafa Raja: Oh, sorry, one more thing, let me…

592 00:43:45.570 ⇒ 00:43:50.230 Mustafa Raja: Let me open up my linear… Give me a moment.

593 00:43:54.550 ⇒ 00:43:55.300 Mustafa Raja: Okay.

594 00:43:56.980 ⇒ 00:44:01.119 Mustafa Raja: And then implement eval… this is eval logging, right?

595 00:44:04.230 ⇒ 00:44:09.620 Samuel Roberts: Implement evals? Yeah, yeah. And that was also the evals we were doing, I think, you know.

596 00:44:09.860 ⇒ 00:44:10.320 Casie Aviles: Yay.

597 00:44:10.320 ⇒ 00:44:14.299 Samuel Roberts: just pick the… The, the analysis we did.

598 00:44:16.040 ⇒ 00:44:23.970 Casie Aviles: And have we, like, selected, the scorers already that we want to… or, like, the metrics we want to score?

599 00:44:26.890 ⇒ 00:44:27.740 Casie Aviles: Or is there…

600 00:44:27.740 ⇒ 00:44:28.309 Samuel Roberts: That’s something.

601 00:44:28.310 ⇒ 00:44:28.890 Casie Aviles: bed.

602 00:44:29.290 ⇒ 00:44:29.740 Mustafa Raja: I think…

603 00:44:29.740 ⇒ 00:44:32.070 Samuel Roberts: Yeah, well, this is… Alright.

604 00:44:32.410 ⇒ 00:44:37.110 Mustafa Raja: Yeah, I think what we have right now is latency, Nlm as a judge?

605 00:44:38.050 ⇒ 00:44:39.349 Casie Aviles: Okay, okay.

606 00:44:40.030 ⇒ 00:44:46.060 Casie Aviles: Yeah, so that should come first, right, before the logging, or am I… Getting it wrong.

607 00:44:46.060 ⇒ 00:44:50.170 Samuel Roberts: It’s kind of… it’s kind of the same. They’re both going to be logged, and… scored.

608 00:44:50.170 ⇒ 00:44:51.320 Casie Aviles: Okay. So, like, they’re… Okay.

609 00:44:51.320 ⇒ 00:44:52.279 Samuel Roberts: They’re the world fair.

610 00:44:53.950 ⇒ 00:44:59.290 Casie Aviles: The other thing was also just, like, running some of the evals on the previous data to compare Mastra to N8N.

611 00:44:59.610 ⇒ 00:45:01.679 Samuel Roberts: But I think we’re good there.

612 00:45:01.680 ⇒ 00:45:05.419 Mustafa Raja: Yeah, yeah, that’s, like, that… That’s, like, done.

613 00:45:06.010 ⇒ 00:45:09.319 Samuel Roberts: Yeah, exactly, that’s what I’m saying, that was the only other EVAS side of it.

614 00:45:10.520 ⇒ 00:45:11.170 Casie Aviles: Okay.

615 00:45:12.960 ⇒ 00:45:14.180 Mustafa Raja: Let me share my screen.

616 00:45:15.660 ⇒ 00:45:21.009 Mustafa Raja: So, for this ticket, what would be the exact, you know, deliverable here?

617 00:45:23.470 ⇒ 00:45:26.379 Samuel Roberts: That the scores are in place, and then logging… yeah.

618 00:45:26.380 ⇒ 00:45:26.990 Casie Aviles: Yes.

619 00:45:28.440 ⇒ 00:45:31.270 Casie Aviles: We don’t need landfills here anymore, right? Or do we still…

620 00:45:31.510 ⇒ 00:45:33.199 Samuel Roberts: I don’t think… no, we don’t even know, no.

621 00:45:33.730 ⇒ 00:45:37.069 Casie Aviles: Okay, because initially we thought of using ground fuse.

622 00:45:37.500 ⇒ 00:45:38.100 Samuel Roberts: Yeah, but…

623 00:45:38.100 ⇒ 00:45:43.370 Mustafa Raja: So for the live scoder, it’s only going to be, doing the relevancy, right?

624 00:45:45.730 ⇒ 00:45:49.389 Samuel Roberts: The live score will only be… yes, and then we’ll have to run,

625 00:45:49.720 ⇒ 00:45:55.559 Samuel Roberts: So maybe we want to add something about the… Like, weekly re-running stuff.

626 00:45:56.020 ⇒ 00:46:03.220 Mustafa Raja: Yeah, so this ticket, rather, is having the scorer ready for weekly cadence, right?

627 00:46:04.040 ⇒ 00:46:04.710 Samuel Roberts: Yes.

628 00:46:05.190 ⇒ 00:46:07.000 Mustafa Raja: Okay, yeah, I just wanted to make sure.

629 00:46:07.750 ⇒ 00:46:08.290 Samuel Roberts: Yep.

630 00:46:08.460 ⇒ 00:46:14.599 Mustafa Raja: Yeah, this is, this is, this is good, then. I think, I, I already shared,

631 00:46:15.570 ⇒ 00:46:17.020 Mustafa Raja: File with you, right?

632 00:46:19.970 ⇒ 00:46:21.880 Samuel Roberts: the files…

633 00:46:22.790 ⇒ 00:46:23.919 Mustafa Raja: For the CSV.

634 00:46:24.580 ⇒ 00:46:26.120 Samuel Roberts: Oh, yes, yes, yes, in CSV.

635 00:46:27.200 ⇒ 00:46:27.810 Mustafa Raja: Yeah.

636 00:46:30.260 ⇒ 00:46:30.920 Samuel Roberts: Yep.

637 00:46:31.020 ⇒ 00:46:32.609 Mustafa Raja: Is it good to move, then?

638 00:46:35.350 ⇒ 00:46:40.569 Samuel Roberts: Yeah, I think that one’s fine.

639 00:46:42.540 ⇒ 00:46:49.329 Samuel Roberts: I’m wondering if we need to add some kind of R&S, though, for the weekly, like, evals, but we can probably wait on that until we have it all set up and…

640 00:46:49.690 ⇒ 00:46:51.060 Mustafa Raja: Yeah, it’s not even deployed, right?

641 00:46:51.060 ⇒ 00:46:55.920 Samuel Roberts: Yeah, exactly, so that’s not working with that. That’s not too over-optimized.

642 00:46:57.200 ⇒ 00:46:59.510 Mustafa Raja: No, don’t… Complete.

643 00:47:19.980 ⇒ 00:47:20.840 Mustafa Raja: Are you?

644 00:47:21.780 ⇒ 00:47:22.470 Samuel Roberts: No.

645 00:47:22.980 ⇒ 00:47:30.270 Mustafa Raja: Oh… Do we want to keep this until, honey.

646 00:47:31.070 ⇒ 00:47:36.979 Mustafa Raja: Yeah, do we want to keep this until we move everything, in NHL? Yes.

647 00:47:37.830 ⇒ 00:47:38.649 Samuel Roberts: Correct, yep.

648 00:47:38.650 ⇒ 00:47:39.240 Mustafa Raja: Okay.

649 00:47:41.140 ⇒ 00:47:44.379 Mustafa Raja: Okay, okay, yeah, this is good, then my board looks good then.

650 00:47:46.120 ⇒ 00:47:46.820 Samuel Roberts: Okay.

651 00:47:48.500 ⇒ 00:47:49.969 Amber Lin: Casey, how do you feel?

652 00:47:50.860 ⇒ 00:47:53.770 Amber Lin: Do you have enough time to finish all of them?

653 00:47:54.120 ⇒ 00:48:00.480 Amber Lin: I know you have also tasks on other clients, like, I… I’m worried about… You taking on too much.

654 00:48:02.370 ⇒ 00:48:12.230 Casie Aviles: For… well, I don’t have much to work on for Eden, and then for Lilo, I have, like, I guess one or two tickets there.

655 00:48:12.750 ⇒ 00:48:14.810 Samuel Roberts: I think I should be able to…

656 00:48:15.460 ⇒ 00:48:19.019 Samuel Roberts: I would say focus your efforts here, and if you need me to grab stuff on Leela, let me know.

657 00:48:19.920 ⇒ 00:48:20.700 Casie Aviles: Okay.

658 00:48:20.810 ⇒ 00:48:25.269 Casie Aviles: Yeah, because the error one is a bit urgent, since…

659 00:48:25.770 ⇒ 00:48:27.150 Samuel Roberts: Yeah.

660 00:48:27.220 ⇒ 00:48:31.200 Casie Aviles: Yvette and Steven were testing that, so I logged immediately.

661 00:48:31.810 ⇒ 00:48:33.800 Casie Aviles: What they were testing with.

662 00:48:34.090 ⇒ 00:48:41.729 Casie Aviles: And I also diagnosed, like, what are the issues which are below in the comment thread, you know?

663 00:48:42.540 ⇒ 00:48:47.280 Casie Aviles: So, I wanna get this to them… Today, at least.

664 00:48:47.580 ⇒ 00:48:48.380 Casie Aviles: Yeah.

665 00:48:50.420 ⇒ 00:48:52.679 Casie Aviles: But also, I might request

666 00:48:52.920 ⇒ 00:48:59.059 Casie Aviles: someone from the team, maybe Mustafa, to also kind of double-check, or, like, just run the tests again.

667 00:48:59.060 ⇒ 00:48:59.700 Mustafa Raja: Yep.

668 00:48:59.700 ⇒ 00:49:13.370 Casie Aviles: Because, like, one of the main things we’re seeing is that it’s… there’s, like, an inconsistency where one person would get the right answer, and then the other would not get the right answer, even if it’s the same…

669 00:49:14.060 ⇒ 00:49:15.610 Casie Aviles: input, right? So…

670 00:49:15.610 ⇒ 00:49:16.320 Mustafa Raja: Yep.

671 00:49:17.230 ⇒ 00:49:18.409 Mustafa Raja: Yeah, so when they get…

672 00:49:18.410 ⇒ 00:49:20.450 Samuel Roberts: Doing lookups by departments and stuff?

673 00:49:22.330 ⇒ 00:49:23.760 Mustafa Raja: No, we’re not doing that.

674 00:49:25.360 ⇒ 00:49:26.920 Samuel Roberts: Oh, I thought we had the router.

675 00:49:27.450 ⇒ 00:49:31.879 Mustafa Raja: Yeah, we removed that. What happens now is…

676 00:49:32.600 ⇒ 00:49:38.229 Mustafa Raja: And you would rather, in the message, mention the departments that it did reference, and then… Right.

677 00:49:38.230 ⇒ 00:49:38.650 Samuel Roberts: Right, right.

678 00:49:38.650 ⇒ 00:49:47.799 Mustafa Raja: If, if the person tells them specifically, search this one, this department for my answer, then it would, you know, look into that.

679 00:49:48.180 ⇒ 00:50:02.399 Amber Lin: I have a question, I’ve been wondering about this for a long time. Like, if we have a UI, why are we still using, like, tech-to-SQL when, like, for very simple questions, they could just use the filters?

680 00:50:02.600 ⇒ 00:50:17.229 Amber Lin: Like, why… why are we… are we able to give them access to the UI in our new, like, Manstra? Because we’re hosting that page ourselves, right? Are we able to just show them that with view-only access? .

681 00:50:17.230 ⇒ 00:50:22.179 Samuel Roberts: We could, yeah, but that would kind of defeat, like, getting out… getting out of the chat and everything, but yeah, you’re right, that’s not a…

682 00:50:23.340 ⇒ 00:50:25.890 Samuel Roberts: Not a bad idea, if that’s what they’re interested in.

683 00:50:26.020 ⇒ 00:50:39.470 Amber Lin: Because, like, I think we’re having a lot of accuracy and completeness issues, and, like, we’re just… like, Andy is just a middleman of the DB and this, and if our DB is good enough…

684 00:50:39.910 ⇒ 00:50:48.910 Amber Lin: like, can’t we just have them search it up? Because then they wouldn’t even have to type the question in anymore, they would just filter.

685 00:50:49.160 ⇒ 00:51:01.489 Amber Lin: like, if we can have this in Andy, like, if we can have it in our new, like, Andy interface without switching tabs, I think it doesn’t… it won’t defeat the purpose of having, like, a chat.

686 00:51:01.780 ⇒ 00:51:02.800 Amber Lin: like, going.

687 00:51:02.800 ⇒ 00:51:03.170 Samuel Roberts: I know.

688 00:51:03.170 ⇒ 00:51:05.639 Amber Lin: It will still be in Andy. Does that make sense?

689 00:51:06.450 ⇒ 00:51:10.870 Samuel Roberts: Not completely. So you’re… you want… I see. …in the Google Chat, like, what…

690 00:51:10.900 ⇒ 00:51:17.070 Amber Lin: Oh, sorry, like, I was thinking for the new master agent, is it still gonna be in Google Chat?

691 00:51:17.070 ⇒ 00:51:17.820 Mustafa Raja: Yeah.

692 00:51:17.820 ⇒ 00:51:18.550 Samuel Roberts: Yeah, yeah.

693 00:51:18.780 ⇒ 00:51:19.110 Casie Aviles: Yeah.

694 00:51:19.110 ⇒ 00:51:21.810 Amber Lin: I see, so it will be a new interface.

695 00:51:23.180 ⇒ 00:51:24.160 Samuel Roberts: No, no.

696 00:51:24.160 ⇒ 00:51:24.930 Casie Aviles: I mean…

697 00:51:24.930 ⇒ 00:51:30.119 Samuel Roberts: I mean, at some point, we could still do that for them, but that wasn’t really part of this. It was just migrating the logic over.

698 00:51:30.720 ⇒ 00:51:31.450 Casie Aviles: Yeah.

699 00:51:31.450 ⇒ 00:51:31.990 Amber Lin: I see.

700 00:51:31.990 ⇒ 00:51:32.500 Casie Aviles: So…

701 00:51:32.770 ⇒ 00:51:41.989 Amber Lin: Yeah, I do want to ask them, what do you guys think? I feel like we’re having so much issues when we already have the DB, but now we’re fixing all, like, the SQL stuff.

702 00:51:42.300 ⇒ 00:51:42.660 Mustafa Raja: Yeah.

703 00:51:44.340 ⇒ 00:51:45.080 Samuel Roberts: Damn.

704 00:51:45.220 ⇒ 00:51:54.080 Mustafa Raja: I think the UI… having the UI in terms of… having the UI would really help, because, we could get rid of the…

705 00:51:54.480 ⇒ 00:51:58.680 Mustafa Raja: Logic that we are doing for departments, specifically.

706 00:51:58.780 ⇒ 00:52:01.380 Mustafa Raja: Because, by,

707 00:52:01.550 ⇒ 00:52:07.119 Mustafa Raja: With that, we could just have a drop-down, they could select the department and ask the question.

708 00:52:07.230 ⇒ 00:52:08.460 Mustafa Raja: So, yeah.

709 00:52:09.010 ⇒ 00:52:12.510 Mustafa Raja: The idea of having the UI might help us or not.

710 00:52:13.300 ⇒ 00:52:18.080 Samuel Roberts: I mean, I was talking to that before, too, when I was… I wanted, like, a better UI in general, and…

711 00:52:18.330 ⇒ 00:52:22.859 Samuel Roberts: you know, not in Google Chat, but that wasn’t part of the scope, so I didn’t.

712 00:52:23.140 ⇒ 00:52:24.120 Amber Lin: Yeah.

713 00:52:24.360 ⇒ 00:52:28.699 Amber Lin: Now, who do we need to talk to about that decision? Because, like…

714 00:52:29.190 ⇒ 00:52:38.119 Amber Lin: we’re just… I don’t want the event to come back and say, like, hey, this doesn’t work, because they’re testing it, and then we just got quite a few feedback that…

715 00:52:38.320 ⇒ 00:52:41.270 Amber Lin: Because it’s not working, we’re seeing issues, so I’m just…

716 00:52:41.500 ⇒ 00:52:53.000 Amber Lin: Just, like, we have all the right things, and, like, I feel like it’s not fair for… for us to get all… all of that stuff when, like, they… they’re just too complicated, they have too much stuff.

717 00:52:53.790 ⇒ 00:52:58.979 Casie Aviles: It’s true, like, I’ve… I think, like, the main challenge here is…

718 00:52:59.090 ⇒ 00:53:07.709 Casie Aviles: being able to stay in sync with all the spreadsheets and all their custom definitions and all of that, and I’ve outlined here, like, what’s…

719 00:53:07.960 ⇒ 00:53:16.700 Casie Aviles: Causing that, because… For example, like, something is named, you know, search key inconsistencies where

720 00:53:17.390 ⇒ 00:53:23.580 Casie Aviles: I mean, I already did this, but then it… Yeah, the SQL generation…

721 00:53:23.850 ⇒ 00:53:26.670 Casie Aviles: I guess my instruction was not very good, so…

722 00:53:27.130 ⇒ 00:53:29.460 Casie Aviles: That’s another layer of complexity there.

723 00:53:30.150 ⇒ 00:53:35.839 Casie Aviles: As opposed to just checking the DB directly, and if they were able to, like, update this

724 00:53:36.290 ⇒ 00:53:38.730 Casie Aviles: Or just use the UI

725 00:53:38.910 ⇒ 00:53:43.830 Casie Aviles: to see everything, and then manage it themselves, I think that would really help.

726 00:53:45.630 ⇒ 00:53:51.930 Casie Aviles: That’s also what I was mentioning to Janiece, where she might become a bottleneck if she’s the only person managing it.

727 00:53:51.930 ⇒ 00:53:52.420 Samuel Roberts: Right.

728 00:53:52.420 ⇒ 00:53:57.990 Casie Aviles: And then… Everyone’s gonna be saying, oh, it’s in the sheets, oh, it’s in the sheets, but…

729 00:53:58.160 ⇒ 00:54:05.619 Casie Aviles: We have to update that then, and then also, like, how would she know if she needs to update it, right?

730 00:54:05.930 ⇒ 00:54:10.770 Samuel Roberts: Do we have clarity on, like, where this… lives in general, like, where…

731 00:54:10.770 ⇒ 00:54:13.220 Casie Aviles: Oh, yeah, yeah, I did, I did ask her.

732 00:54:13.350 ⇒ 00:54:18.199 Casie Aviles: And what she told me is that they really are relying on the sheets.

733 00:54:19.520 ⇒ 00:54:27.849 Casie Aviles: And some sheets are… and what makes this more complex is some sheets don’t really change much.

734 00:54:28.160 ⇒ 00:54:33.549 Casie Aviles: And then some has, like, have… has a different pace where they get updated every month, you know?

735 00:54:34.580 ⇒ 00:54:35.090 Samuel Roberts: So…

736 00:54:35.090 ⇒ 00:54:35.780 Amber Lin: No.

737 00:54:36.350 ⇒ 00:54:38.329 Samuel Roberts: They are, so the sheets is the source of truth.

738 00:54:38.830 ⇒ 00:54:46.220 Amber Lin: Yeah, and different people update the sheets, and different department owns the sheets, and then they have, like…

739 00:54:46.400 ⇒ 00:55:01.079 Amber Lin: one sheet for each department, they’re formatted differently, so we can’t really just sync it to our system, like, we have to manually be told what has changed, because it might just be one line, or one cell that has changed. So…

740 00:55:01.260 ⇒ 00:55:06.759 Amber Lin: Wait… If they’re used to sheets, then this is what they should use.

741 00:55:07.600 ⇒ 00:55:13.649 Samuel Roberts: I was gonna say, like, that’s… I mean, the real sell here is getting them into something that is gonna be more standardized and…

742 00:55:13.650 ⇒ 00:55:14.410 Amber Lin: Yeah.

743 00:55:14.410 ⇒ 00:55:16.380 Samuel Roberts: Like, that’s a process thing on their end.

744 00:55:16.930 ⇒ 00:55:17.779 Samuel Roberts: Help them with.

745 00:55:18.120 ⇒ 00:55:34.800 Amber Lin: Because ultimately, they want their service managers, which is the ones that’s actually go, like, managing the inspectors and technicians to update this, because right now, they have to tell Janice, and Janice goes, tells people who own the spreadsheets to update it. Like, there’s… there’s a…

746 00:55:34.940 ⇒ 00:55:35.460 Samuel Roberts: That’s…

747 00:55:35.460 ⇒ 00:55:36.210 Amber Lin: Huh.

748 00:55:36.210 ⇒ 00:55:43.949 Samuel Roberts: Something we could pitch them on is… is, you know, this could be a thing that different people could log into, and keep this up to date, and make this

749 00:55:44.410 ⇒ 00:55:52.390 Samuel Roberts: completely supplant sheets. That might take a little more, or definitely will take a little more work to, like, polish it up and add all the edge cases and everything.

750 00:55:53.970 ⇒ 00:55:55.290 Samuel Roberts: But, I think…

751 00:55:55.440 ⇒ 00:56:00.309 Samuel Roberts: For their purposes, getting off of just updating sheets and having everyone have to rely on that.

752 00:56:00.610 ⇒ 00:56:05.040 Samuel Roberts: Seems like a better bet. And then we could also give read access and things like that.

753 00:56:05.270 ⇒ 00:56:21.399 Amber Lin: Yeah. I think to summarize, there’s two main things we’re aiming here. It’s like, one is to get the people who update she, so get the source of truth in here. Like, that’s no matter what we do with Andy, with the chat and all that, we need them to.

754 00:56:21.400 ⇒ 00:56:21.900 Samuel Roberts: update this.

755 00:56:21.900 ⇒ 00:56:38.149 Amber Lin: So, that’s gonna happen regardless, like, we want to get this standardized. And two, it’s, like, we’re right now talking about, do we need Andy to answer the zip code questions? Because then, it’s like, how… how much tweaking will it take, and will we ever be perfect?

756 00:56:38.150 ⇒ 00:56:44.970 Amber Lin: And, like, how much are we really losing out on just minor wording conversion issues that

757 00:56:44.970 ⇒ 00:56:49.930 Amber Lin: Like, is that really fixable, or how much are we losing out because of that? Like, what do you guys think?

758 00:56:52.290 ⇒ 00:56:53.290 Samuel Roberts: Boom.

759 00:56:56.100 ⇒ 00:56:57.230 Samuel Roberts: I don’t know, I think…

760 00:56:58.230 ⇒ 00:57:04.369 Samuel Roberts: There’s… there’s some things, like, even like this, like, inconsistencies that we can get over time and… and get fixed, but, like.

761 00:57:05.330 ⇒ 00:57:11.849 Samuel Roberts: I don’t know… How to fix the source of truth issue without them Committing to this, you know?

762 00:57:13.850 ⇒ 00:57:19.589 Samuel Roberts: That’s the bigger one in my mind. Like, I bet we can… we can squash all these little things here over time.

763 00:57:19.890 ⇒ 00:57:20.910 Amber Lin: But, like… Hmm.

764 00:57:21.090 ⇒ 00:57:24.849 Samuel Roberts: It doesn’t matter if it knows estimator versus inspector, if the data’s just old.

765 00:57:25.350 ⇒ 00:57:41.989 Amber Lin: I see. So, what do you think is the biggest barrier? Is it… like, do we need enough people to use it for them to update this, or do we need to… like, are they not updating it because they’ve seen so many errors and they don’t trust Andy enough? Like, what’s the barrier?

766 00:57:41.990 ⇒ 00:57:42.420 Samuel Roberts: Alright.

767 00:57:42.420 ⇒ 00:57:43.040 Amber Lin: Is…

768 00:57:43.250 ⇒ 00:57:50.230 Samuel Roberts: Yeah, I don’t know, I mean, I don’t have a ton of insight into, like, how they’re using it that way, so I would default to someone else answering that.

769 00:57:50.230 ⇒ 00:57:51.840 Casie Aviles: You mean the admin tool?

770 00:57:52.960 ⇒ 00:57:53.510 Amber Lin: Yeah.

771 00:57:53.670 ⇒ 00:57:58.850 Amber Lin: Like, them using this as source of truth to update this, because it’s just Janiece right now.

772 00:57:58.990 ⇒ 00:58:10.649 Casie Aviles: Yeah, it’s just Janiece right now. I don’t think it’s been released to other people, and I think that’s… that’s what I believe they were also mentioning, where they don’t want it released yet to many… a lot of…

773 00:58:11.450 ⇒ 00:58:17.280 Casie Aviles: users, just them for now. But I think they’d also plan to loop in more.

774 00:58:17.970 ⇒ 00:58:24.059 Casie Aviles: people, if it… if it works well. I mean, it doesn’t… hasn’t really failed. I mean, I haven’t really seen, like.

775 00:58:24.900 ⇒ 00:58:31.020 Casie Aviles: Any issues yet, but then again, Jenny’s hasn’t used it much yet, so…

776 00:58:31.020 ⇒ 00:58:31.760 Samuel Roberts: Yeah.

777 00:58:31.760 ⇒ 00:58:41.620 Amber Lin: Yeah, I feel like it will be hard to get them to use this just as a updating tool, because if it’s just an updating tool for Andy, then

778 00:58:41.620 ⇒ 00:58:56.709 Amber Lin: all they’re caring about is if Andy is correct, and right now it’s Andy is not correct. But, like, that’s why I wanted us to see if we can get CSRs to just use this UI, because then everybody will be using it, and they’ll be saying, like, hey, this is…

779 00:58:56.760 ⇒ 00:59:03.720 Amber Lin: this is not right, like, I searched this up, this occurred, but then I, like, I wouldn’t know where to add in

780 00:59:03.930 ⇒ 00:59:08.720 Amber Lin: errors if they see it, because this is not Andy anymore, like, I just don’t know.

781 00:59:11.730 ⇒ 00:59:23.710 Amber Lin: Like, I feel like getting this as a source of truth depends on a lot of usage, and right now, usage is blocked because we can’t get any right, because there’s just so many intricacies.

782 00:59:23.710 ⇒ 00:59:30.399 Samuel Roberts: Well, I think there’s also some other issues here where, like, I’ll see things coming through, and it’ll be like, so-and-so told me this is the actual truth.

783 00:59:30.650 ⇒ 00:59:33.069 Samuel Roberts: And I’ll be like, well, then what is it? Is it…

784 00:59:33.870 ⇒ 00:59:40.159 Samuel Roberts: Like, where… like, I just not… I think the bigger issue here is, like, the central docs as well, and the spreadsheets, like…

785 00:59:41.550 ⇒ 00:59:44.270 Samuel Roberts: they need… a better…

786 00:59:45.650 ⇒ 00:59:53.050 Samuel Roberts: way to know what the truth is, I guess, in general. Like, that, I feel like, is the thing that, like, we can’t just solve for them either, we need to work with them to…

787 00:59:55.880 ⇒ 00:59:57.490 Samuel Roberts: Either… That makes sense.

788 00:59:57.620 ⇒ 01:00:03.439 Samuel Roberts: like, because I worry that people are just like, that’s not right, I heard that so-and-so’s not doing that anymore, and I’m like…

789 01:00:03.970 ⇒ 01:00:07.149 Samuel Roberts: I see. Like, where is that coming from? Where is that in here?

790 01:00:07.150 ⇒ 01:00:13.809 Amber Lin: I think they’re willing to collaborate, but they just don’t… they also don’t know. They never got really instruction from us, either.

791 01:00:13.810 ⇒ 01:00:17.869 Samuel Roberts: Right, that’s what I’m saying, I think we need to really help them with that, and that’s probably a bigger, like…

792 01:00:18.720 ⇒ 01:00:24.469 Samuel Roberts: Bigger project to, like, really… Audit all of that stuff, and…

793 01:00:25.600 ⇒ 01:00:33.899 Samuel Roberts: I mean, the databases was kind of the first swing at that, I felt like, and if it’s not… if no one’s using it to update and keep it up to date, and the sheets are still the source of truth.

794 01:00:35.100 ⇒ 01:00:42.119 Samuel Roberts: I don’t know what else to push them to. I mean, unless you want to give them read-only access, and then they can just try to use that outside Andy.

795 01:00:42.270 ⇒ 01:00:47.210 Samuel Roberts: And just to confirm it, but that seems like it’s still not getting people to trust Andy.

796 01:00:48.710 ⇒ 01:00:49.690 Amber Lin: Yeah…

797 01:00:50.780 ⇒ 01:01:05.170 Amber Lin: I mean, is… is there… then, I guess, taking a step back, do we have to do text-to-SQL, or is there any way, to give them, like, there’s 1, 2, 3, 4, 5, 6, 7, like, 7 fields?

798 01:01:05.290 ⇒ 01:01:16.259 Amber Lin: Like, is there any way to give them a way to, like, enter in… like, we’ll give them set selections in the fields? Like, is there any way to let them pick that?

799 01:01:16.790 ⇒ 01:01:20.500 Amber Lin: Like, is there a selection tool in Google Chat that we can use?

800 01:01:20.500 ⇒ 01:01:22.569 Samuel Roberts: And Google Chat, I don’t… I don’t know.

801 01:01:23.200 ⇒ 01:01:28.329 Amber Lin: Or, like, have them do, like, pick one option, like, reply with…

802 01:01:28.330 ⇒ 01:01:28.829 Casie Aviles: Yeah, no, no.

803 01:01:28.830 ⇒ 01:01:35.530 Amber Lin: one of the options here. Like, we have to standardize it somehow, like, and we can’t rely on them giving us the right…

804 01:01:36.260 ⇒ 01:01:39.070 Amber Lin: Like, we have to give them the standardized text.

805 01:01:39.200 ⇒ 01:01:44.260 Amber Lin: Like, they’re not gonna give that to us. They’re always gonna say sales inspectors, like.

806 01:01:44.260 ⇒ 01:01:44.730 Casie Aviles: Estimated.

807 01:01:47.420 ⇒ 01:01:48.030 Casie Aviles: I mean…

808 01:01:48.030 ⇒ 01:01:50.129 Samuel Roberts: That, that I’m less worried about.

809 01:01:50.240 ⇒ 01:01:54.390 Samuel Roberts: I feel like they can… we can understand those things and… and give that into the…

810 01:01:54.520 ⇒ 01:01:57.009 Samuel Roberts: the prompt, right? Maybe I’m wrong, Casey, I don’t know.

811 01:01:57.480 ⇒ 01:02:03.600 Casie Aviles: Yeah, that’s what I’ve been doing, like, whenever they say something differently, I have to update

812 01:02:03.950 ⇒ 01:02:11.940 Casie Aviles: the instructions, but then that’s, also something I have to do each time I find a new edge case, or, like.

813 01:02:12.180 ⇒ 01:02:15.510 Casie Aviles: They, they have a new…

814 01:02:15.670 ⇒ 01:02:23.689 Casie Aviles: Way of calling estimators, you know, or inspectors, since… the data in… Andy is kind of…

815 01:02:24.140 ⇒ 01:02:27.740 Casie Aviles: I mean, the columns or, like, the values there are…

816 01:02:28.810 ⇒ 01:02:32.519 Casie Aviles: Just as is, I… they don’t… yeah, sorry, I might be…

817 01:02:32.760 ⇒ 01:02:37.460 Casie Aviles: That might be, rambly or unclear, but…

818 01:02:37.460 ⇒ 01:02:39.579 Samuel Roberts: No, I think I get it. Yeah, I…

819 01:02:47.860 ⇒ 01:02:54.209 Amber Lin: I mean, I think Google Chat has the selection input, like, there’s, like, selection menus.

820 01:02:54.210 ⇒ 01:02:59.889 Samuel Roberts: So, the idea is, like, they would say what, like… Like, I need…

821 01:03:01.160 ⇒ 01:03:06.740 Samuel Roberts: To look up something, and then it would just pop up with a thing that they can… Yeah.

822 01:03:06.740 ⇒ 01:03:10.219 Casie Aviles: Maybe a slash command or something, and then…

823 01:03:10.220 ⇒ 01:03:11.280 Samuel Roberts: Yeah, we could look at that.

824 01:03:11.280 ⇒ 01:03:12.200 Casie Aviles: I could probably build.

825 01:03:12.200 ⇒ 01:03:12.780 Amber Lin: Yeah.

826 01:03:13.680 ⇒ 01:03:23.359 Casie Aviles: a custom… I think Google Chat has, like, we can build something custom, like what we did for the thumbs up, thumbs down feedback.

827 01:03:23.950 ⇒ 01:03:25.420 Samuel Roberts: Right, right.

828 01:03:25.420 ⇒ 01:03:30.939 Casie Aviles: Yeah, we’ll have to build that again, and then, which essentially renders the text to SQL.

829 01:03:31.290 ⇒ 01:03:33.060 Casie Aviles: Useless.

830 01:03:33.060 ⇒ 01:03:39.130 Amber Lin: Yeah, I’m trying to, like, bypass the text-to-SQL, like, that’s what I see what we’ll.

831 01:03:39.130 ⇒ 01:03:42.479 Samuel Roberts: Is the text-to-SQL the issue, though? Like…

832 01:03:43.790 ⇒ 01:03:44.120 Amber Lin: Have we

833 01:03:44.530 ⇒ 01:03:56.370 Amber Lin: And, like, I think text-to-SQL is less of an issue than the hundred different ways that they ask things, because they have a hundred different people asking things in different ways.

834 01:03:56.940 ⇒ 01:04:06.390 Casie Aviles: And yeah, and, like, the data that they have is in so many different formats, and we’re not, like, getting everything that they need, because

835 01:04:06.900 ⇒ 01:04:15.189 Casie Aviles: Once, sometimes I’ve added something, and then I just find out, oh, there’s something missing there, it’s not, it should be in the sheets, it should be in the sheets, and then…

836 01:04:15.920 ⇒ 01:04:21.499 Casie Aviles: Oh, yeah, like, there’s, there’s, there are, like, new fields, you know, that they, they have, or, like…

837 01:04:22.630 ⇒ 01:04:25.159 Casie Aviles: So that’s kind of, like, the challenge there.

838 01:04:28.280 ⇒ 01:04:28.690 Samuel Roberts: Right.

839 01:04:28.690 ⇒ 01:04:35.589 Casie Aviles: it’s not standardized. There’s a lot of different input, and I’m not sure, like, what’s the best way we can

840 01:04:36.360 ⇒ 01:04:44.920 Casie Aviles: Unless they use this tool as, like, the only tool to… You know, consolidate everything.

841 01:04:46.120 ⇒ 01:04:49.929 Samuel Roberts: I think that’s… that’s… that’s really the answer here, is that they need to have…

842 01:04:50.380 ⇒ 01:04:57.550 Samuel Roberts: One place, one source of truth, and then we can work around some of those edge cases, but even if we just get everyone using the same thing.

843 01:04:57.910 ⇒ 01:04:59.219 Samuel Roberts: Is a big win.

844 01:04:59.670 ⇒ 01:05:03.789 Samuel Roberts: Because, like… We can add aliases, we can add…

845 01:05:04.580 ⇒ 01:05:09.639 Samuel Roberts: you know, whatever terminology here or there. We can also push them to use it more. We can have

846 01:05:09.750 ⇒ 01:05:12.330 Samuel Roberts: Clarifications that we can add.

847 01:05:15.450 ⇒ 01:05:16.540 Samuel Roberts: But…

848 01:05:19.600 ⇒ 01:05:22.790 Samuel Roberts: I don’t… I don’t know if I have a good…

849 01:05:23.920 ⇒ 01:05:37.400 Amber Lin: So what would be our next steps, or next thing we need to explore? Like, I know we have some things that we’re unsure about, but what would make that answer easier? Like, what would you need to make that decision?

850 01:05:38.170 ⇒ 01:05:48.220 Samuel Roberts: I would need to know how we could get them off the spreadsheets, and if that’s something they’re willing to do, and if, like, making this database more of a robust tool is something they’re interested in.

851 01:05:48.580 ⇒ 01:05:55.339 Samuel Roberts: I kind of assumed that there was another place where they had all of these assignments, and the spreadsheets were just for people to look up stuff, but…

852 01:05:55.560 ⇒ 01:05:57.509 Samuel Roberts: I guess that is really…

853 01:05:57.510 ⇒ 01:05:58.010 Amber Lin: willing.

854 01:05:58.010 ⇒ 01:06:00.269 Samuel Roberts: managing everything? Yeah, that’s crazy, okay.

855 01:06:01.020 ⇒ 01:06:01.670 Amber Lin: Yeah.

856 01:06:02.000 ⇒ 01:06:15.549 Amber Lin: Do we also need a better understanding of what type of error? Like, what is really the true cause of problem? Like, is it the text to SQL? Is it the analysis? Like, do we have a certain idea of…

857 01:06:15.660 ⇒ 01:06:20.860 Amber Lin: Like, what is the biggest issue, so that we can… Make a decision about it.

858 01:06:24.800 ⇒ 01:06:29.239 Samuel Roberts: I don’t, personally, but I’m not in this… this issue as much.

859 01:06:30.030 ⇒ 01:06:30.960 Casie Aviles: Okay.

860 01:06:30.960 ⇒ 01:06:31.590 Samuel Roberts: Okay, so…

861 01:06:31.590 ⇒ 01:06:37.959 Casie Aviles: I mean, for… just based off what I’m seeing, though, I mean, maybe we need to, like.

862 01:06:38.160 ⇒ 01:06:41.750 Casie Aviles: get really used with Mustafa’s,

863 01:06:42.410 ⇒ 01:06:45.480 Casie Aviles: Produced, you know, with the kinds of errors, but…

864 01:06:46.350 ⇒ 01:06:53.369 Casie Aviles: Just based on what I’m seeing, it’s really, the database and the challenge of

865 01:06:53.760 ⇒ 01:06:57.809 Casie Aviles: Or, like, keeping the database updated, right? That’s, like, the main…

866 01:06:58.660 ⇒ 01:07:05.379 Casie Aviles: issue, I would say, and I believe Sam’s right, we’re in, like, minor wording things.

867 01:07:05.620 ⇒ 01:07:07.050 Casie Aviles: That can be ironed out.

868 01:07:07.050 ⇒ 01:07:07.720 Amber Lin: Hmm.

869 01:07:08.000 ⇒ 01:07:14.179 Casie Aviles: And… based on what… so I’ve also been trying to tag tickets with…

870 01:07:14.350 ⇒ 01:07:17.939 Casie Aviles: New labels, where… is it a zip code issue, or a…

871 01:07:18.500 ⇒ 01:07:24.780 Casie Aviles: Or, like, a zip code update issue, or a text-to-SQL issue, and I can see that there’s just more…

872 01:07:25.270 ⇒ 01:07:27.890 Casie Aviles: outdated zip code.

873 01:07:27.890 ⇒ 01:07:29.900 Amber Lin: Awesome, okay, awesome, okay.

874 01:07:29.900 ⇒ 01:07:30.899 Samuel Roberts: That’s good to know, okay.

875 01:07:31.750 ⇒ 01:07:33.520 Amber Lin: Right, so…

876 01:07:34.670 ⇒ 01:07:50.809 Amber Lin: let’s go have the conversation about the spreadsheets. I would actually like you guys to be there, because, like, I… I’m gonna lose some stuff in translation. We have a meeting on Thursday that I think…

877 01:07:50.960 ⇒ 01:07:58.999 Amber Lin: It’s around, like, probably an hour from now tomorrow? Like, Sam, would you be able to join there, or either?

878 01:07:59.000 ⇒ 01:08:02.680 Samuel Roberts: Yeah, yeah, that’s on my calendar, I just haven’t been going because I wasn’t super involved and stuff.

879 01:08:02.680 ⇒ 01:08:19.629 Amber Lin: Okay, that makes sense. Alright, so I think we should have that conversation, because, like, we’re stuck, they want to do it, but they feel like things have been failing, and they don’t really understand what’s going on. So let’s have that conversation with them tomorrow, and then we’ll figure out

880 01:08:19.689 ⇒ 01:08:27.169 Amber Lin: what we have to do, because you’re right, like, I… like, going away from AI defeats our whole entire purpose.

881 01:08:27.390 ⇒ 01:08:28.130 Amber Lin: So…

882 01:08:28.130 ⇒ 01:08:28.560 Samuel Roberts: Right.

883 01:08:28.560 ⇒ 01:08:33.250 Amber Lin: Cool. Alright, so…

884 01:08:33.490 ⇒ 01:08:41.549 Amber Lin: I like that, and then we’ll have that convo tomorrow. I have a… I started a checklist for the…

885 01:08:41.729 ⇒ 01:08:45.029 Amber Lin: For the different departments.

886 01:08:46.649 ⇒ 01:08:58.110 Amber Lin: I think we’ll need to schedule time to go over that together. Like, I can probably do that with either Casey or Mustafa. So I’ll book time with you guys to go over that.

887 01:09:00.859 ⇒ 01:09:01.529 Casie Aviles: Okay.

888 01:09:01.979 ⇒ 01:09:04.469 Amber Lin: Let me just flash you the…

889 01:09:05.099 ⇒ 01:09:07.639 Amber Lin: I’ll share a screen, like, this is the…

890 01:09:07.869 ⇒ 01:09:10.289 Amber Lin: kind of the rough structure I have, so…

891 01:09:10.439 ⇒ 01:09:19.009 Amber Lin: For each department, looking at zip codes, linking at central docs, looking at… like…

892 01:09:19.609 ⇒ 01:09:27.589 Amber Lin: I feel like this is not a complete list, but then, like, after each department, we’ll also need to look at, like, our overall

893 01:09:27.719 ⇒ 01:09:35.359 Amber Lin: oh, by the ways, cancellations, templates, and all that. So, essentially, the… the Gantt that we have, but…

894 01:09:35.769 ⇒ 01:09:42.929 Amber Lin: we should make a checklist of how… where we’re at, and so the client have… has a better idea. So I need to make…

895 01:09:43.219 ⇒ 01:09:46.989 Amber Lin: this checklist for… Tomorrow?

896 01:09:47.119 ⇒ 01:09:49.999 Amber Lin: So, like…

897 01:09:51.029 ⇒ 01:09:57.719 Amber Lin: Can I grab time with… like, Casey and Mustafa, are you guys free tomorrow? Can I grab time with you to talk about this?

898 01:09:58.980 ⇒ 01:10:00.860 Casie Aviles: Yeah, yeah, I should have time.

899 01:10:01.140 ⇒ 01:10:01.840 Amber Lin: Okay.

900 01:10:02.330 ⇒ 01:10:05.529 Amber Lin: Cool. That’s… that’s all I need.

901 01:10:05.920 ⇒ 01:10:09.080 Amber Lin: I’ll find a time here.

902 01:10:09.910 ⇒ 01:10:13.759 Mustafa Raja: By department, are we going to take a look at, central docs?

903 01:10:14.670 ⇒ 01:10:19.890 Amber Lin: Yeah, I think zip codes, central docs, and how things are operating right now.

904 01:10:20.820 ⇒ 01:10:21.460 Mustafa Raja: Okay.

905 01:10:21.940 ⇒ 01:10:26.429 Amber Lin: Yeah. Maybe we should all be… I’ll invite both of you, then.

906 01:10:26.670 ⇒ 01:10:29.909 Amber Lin: Just to make… make sure we have enough people.

907 01:10:30.160 ⇒ 01:10:32.440 Amber Lin: Alright, yeah, that’s fine.

908 01:10:32.440 ⇒ 01:10:33.860 Mustafa Raja: How are we setting up this one?

909 01:10:34.820 ⇒ 01:10:35.490 Amber Lin: Pardon me?

910 01:10:35.800 ⇒ 01:10:37.479 Mustafa Raja: Sorry, when are we setting this meeting?

911 01:10:37.870 ⇒ 01:10:44.030 Amber Lin: Probably tomorrow, does that work for you guys? Like, right after… after the Eden stand-up?

912 01:10:44.470 ⇒ 01:10:45.940 Mustafa Raja: Wait, let me see, let me see.

913 01:10:46.440 ⇒ 01:10:47.210 Mustafa Raja: It says you’re.

914 01:10:47.210 ⇒ 01:10:48.679 Amber Lin: street, but I’m not sure.

915 01:10:48.680 ⇒ 01:10:51.269 Mustafa Raja: Yeah, because I have, defaulted…

916 01:10:51.520 ⇒ 01:10:54.790 Mustafa Raja: Sync. That is not on my calendar right now.

917 01:10:55.260 ⇒ 01:10:56.910 Amber Lin: Okay, when would that be?

918 01:10:57.670 ⇒ 01:11:03.390 Mustafa Raja: Let me, let me see, so that, so… yeah.

919 01:11:03.550 ⇒ 01:11:07.520 Mustafa Raja: That’s an hour after the Eden Daily Stand Up.

920 01:11:08.070 ⇒ 01:11:10.749 Amber Lin: Gotcha, so that would be… so I’ll squeeze it in here.

921 01:11:10.750 ⇒ 01:11:11.749 Mustafa Raja: Yeah. Like that.

922 01:11:11.750 ⇒ 01:11:12.380 Amber Lin: Okay.

923 01:11:17.690 ⇒ 01:11:27.120 Amber Lin: Great. And then I have the conversation with the… with the, like, Friday as well. Cool. That’s all, that’s great. This was really productive, thank you guys.

924 01:11:27.120 ⇒ 01:11:29.160 Samuel Roberts: Yes, yes, this was. I’m glad.

925 01:11:29.160 ⇒ 01:11:31.060 Amber Lin: Awesome. Alright, bye guys.

926 01:11:31.060 ⇒ 01:11:31.830 Mustafa Raja: Thank you.

927 01:11:31.830 ⇒ 01:11:32.480 Samuel Roberts: Bye, everyone.