Meeting Title: ABC Standup Date: 2025-07-18 Meeting participants: Awaish Kumar, Mustafa Raja, Annie Yu, Casie Aviles, Amber Lin


WEBVTT

1 00:05:01.220 00:05:02.350 Amber Lin: Hello!

2 00:05:04.550 00:05:05.780 Annie Yu: Hello!

3 00:05:06.140 00:05:07.440 Amber Lin: Thank you, everybody.

4 00:05:07.910 00:05:15.740 Amber Lin: I’m just gonna check on the linear board, really fast. And then I wanted us to talk a bit about

5 00:05:16.150 00:05:28.944 Amber Lin: the roadmap cause we’re going to add another department before we were just helping the pest department. And now we’re gonna also add mechanical. And then I think that brings a few challenges on

6 00:05:30.020 00:05:37.230 Amber Lin: the chat, bot, and how it answers questions. So just as a pre phase, we’ll talk about that.

7 00:05:38.225 00:05:44.630 Amber Lin: I guess, Annie, we’ll talk about this, and then you can feel free to hop off. Would you be? Would you

8 00:05:44.880 00:05:47.136 Amber Lin: take a look at

9 00:05:47.770 00:05:53.460 Annie Yu: Just did, and I still don’t see the added Timestamp. Fields.

10 00:05:53.560 00:05:55.650 Amber Lin: Oh, okay,

11 00:05:57.840 00:06:04.749 Amber Lin: Do you know, I guess Luke is already on vacation

12 00:06:05.254 00:06:09.150 Amber Lin: wish. Do you know what might be the cause of that.

13 00:06:10.700 00:06:15.449 Awaish Kumar: Yeah, I’m not sure, like last time I from us

14 00:06:16.410 00:06:18.729 Awaish Kumar: ask Luke to debug it. So.

15 00:06:18.730 00:06:19.380 Amber Lin: Hmm.

16 00:06:20.360 00:06:27.660 Awaish Kumar: Yeah, I’m not sure he didn’t told me anything, and Yup, I don’t know the status.

17 00:06:29.770 00:06:38.309 Amber Lin: I see, cause he’s going to be. He’s gonna be gone Monday, Tuesday, Wednesday, I believe.

18 00:06:39.122 00:06:40.960 Amber Lin: So we’ll have.

19 00:06:41.510 00:06:43.150 Amber Lin: So yeah, we’ll have.

20 00:06:43.150 00:06:45.740 Amber Lin: I rely on you to help. Okay.

21 00:06:45.740 00:06:47.759 Awaish Kumar: But I can do it on Monday.

22 00:06:48.350 00:06:56.659 Amber Lin: Yeah, totally. No rush. You can see here the cycle still has a week. Just wanna make both of you aware that Luke’s not gonna be here

23 00:06:56.830 00:07:02.519 Amber Lin: for the 1st part next week, so can I help assign this to always? Can I assign it to you?

24 00:07:04.075 00:07:04.580 Awaish Kumar: Yes.

25 00:07:05.160 00:07:12.310 Amber Lin: Okay, okay, sounds good. I also had a task for Luke to set up. Dbt,

26 00:07:13.850 00:07:21.929 Amber Lin: I don’t think that’s that urgent. I’ll probably ask him to do it once he’s back. Don’t want to take up too much a wish of your time.

27 00:07:22.100 00:07:26.169 Amber Lin: so I’ll I’ll ask him to do these 2 when he’s back.

28 00:07:26.840 00:07:27.910 Amber Lin: Okay?

29 00:07:28.100 00:07:34.350 Amber Lin: Think that’s that’s that, Annie. I know you’re blocked by that. So I’ll just.

30 00:07:35.180 00:07:38.399 Amber Lin: I’ll mark it as blocked. And then

31 00:07:39.473 00:07:44.180 Amber Lin: any other things you see for the dashboard. That’s notable. You want to bring up.

32 00:07:45.748 00:07:48.420 Annie Yu: Not at the moment.

33 00:07:49.570 00:07:50.210 Amber Lin: Okay.

34 00:07:50.640 00:07:57.889 Amber Lin: sounds good. Feel free to hop off. I’ll talk about. I just have a few things I want to talk about with. The engineers.

35 00:07:58.270 00:07:58.880 Annie Yu: Okay.

36 00:07:59.270 00:07:59.840 Amber Lin: Okay.

37 00:08:00.070 00:08:01.230 Annie Yu: Thank you. Guys. Thank you.

38 00:08:07.440 00:08:11.220 Amber Lin: Oh, I’m sorry I’m gonna mark this as done. It’s great.

39 00:08:11.220 00:08:17.640 Mustafa Raja: Yeah, I have the the one, too, in the in the review.

40 00:08:18.230 00:08:19.900 Amber Lin: Oh, wait! Where did.

41 00:08:20.030 00:08:22.860 Mustafa Raja: Where did I put it?

42 00:08:22.860 00:08:25.610 Mustafa Raja: And the review from my view, maybe refresh.

43 00:08:25.940 00:08:38.080 Amber Lin: Yeah, let me go check. Maybe it just got didn’t got added to the cycle or got booted out. That one. Yeah. I think it just wasn’t in cycle. That’s all. Okay.

44 00:08:41.970 00:08:46.500 Amber Lin: Hmm, oh, okay.

45 00:08:46.910 00:08:55.789 Amber Lin: sounds good. I will go check this, probably Monday. Did you link that? Okay, you did. Okay, awesome. I’ll go check it.

46 00:08:55.940 00:08:56.515 Amber Lin: Monday.

47 00:08:57.280 00:08:57.840 Mustafa Raja: Yep.

48 00:08:59.710 00:09:05.180 Amber Lin: Currently, yeah, this one.

49 00:09:08.650 00:09:11.869 Amber Lin: I guess. Casey, what are? What are you currently working on?

50 00:09:13.260 00:09:15.170 Casie Aviles: Yeah, so yeah, I’m working on.

51 00:09:15.780 00:09:21.640 Casie Aviles: I think I have it in internal review. But okay, yeah, I see the rag thing. Yeah.

52 00:09:21.640 00:09:28.589 Amber Lin: I see awesome if you have time today. Would you mind just quickly adding the service area spreadsheet to Nan.

53 00:09:30.080 00:09:31.227 Casie Aviles: Oh, okay. Okay.

54 00:09:31.610 00:09:35.160 Amber Lin: It’s also that spreadsheet has already.

55 00:09:35.530 00:09:36.980 Amber Lin: Oh, dear.

56 00:09:37.910 00:09:43.219 Amber Lin: okay, well, I guess we’ll have to talk about that, too. It’s a lot of is that a lot of tokens.

57 00:09:44.420 00:09:52.030 Casie Aviles: 30,000. I think the yeah, the the standard Gpt for open handle. This.

58 00:09:54.020 00:09:54.510 Amber Lin: Okay.

59 00:09:54.900 00:09:55.610 Casie Aviles: Yeah.

60 00:09:56.878 00:10:01.489 Amber Lin: Let me grab that spreadsheet for you, and then I would love us to talk about

61 00:10:01.610 00:10:06.380 Amber Lin: like, how do we route things? I’m just gonna put that one in there

62 00:10:06.680 00:10:17.239 Amber Lin: so just quickly. This spreadsheet is also by zip code. So it wouldn’t be much different than how we routed it before. So this is purpose of

63 00:10:18.561 00:10:22.319 Amber Lin: this spreadsheet is for us.

64 00:10:28.180 00:10:37.170 Amber Lin: Yeah, okay. And then today, I I wanted to ask you guys how you think we can approach

65 00:10:39.670 00:10:42.399 Amber Lin: routing. So I guess the current problem

66 00:10:43.878 00:10:51.260 Amber Lin: is that we’re currently just dumping a massive context. It’s gonna take gonna take a long time.

67 00:10:51.460 00:10:58.369 Amber Lin: The initial rack, as we tested, is a lot faster. But it’s not as specific.

68 00:10:58.930 00:11:10.980 Amber Lin: And then we’re also gonna be adding new departments. So new Csr divisions, and then in the future, maybe we’re even gonna add, like their Hr or other departments. So I wanted to ask you guys what

69 00:11:11.100 00:11:14.209 Amber Lin: what options we see for

70 00:11:15.860 00:11:24.599 Amber Lin: like improving the situation. This is what AI recommended, but I want to hear from you all as a as a group. What would what we think.

71 00:11:29.990 00:11:38.010 Casie Aviles: Yeah, so yeah, just based on what I what I have at the top of my mind.

72 00:11:38.810 00:11:45.299 Casie Aviles: I do think that the the the approach where we do the context thing is

73 00:11:45.850 00:11:53.250 Casie Aviles: I I don’t think it. It’s that scalable, because, you know, especially with what happened just recently with

74 00:11:55.166 00:11:59.129 Casie Aviles: with the increase. The increasing response. Times.

75 00:11:59.480 00:12:05.529 Amber Lin: - yeah, I think context is out of the probably out of the question.

76 00:12:07.281 00:12:11.599 Amber Lin: Cause, even for one department, context is not enough anymore.

77 00:12:12.620 00:12:18.270 Casie Aviles: Yeah, it was. It was fine when we had just a cup, the central dock and the

78 00:12:18.680 00:12:27.079 Casie Aviles: technician spreadsheet. But now we’re getting more. So, yeah, I would, okay, yeah, go ahead.

79 00:12:27.080 00:12:46.539 Mustafa Raja: Yeah, I would want to add that last night I I also checked with the rag implementation that Casey did. I feel the answers now are much better than we last tested it. Maybe we should. We should test it right now to see where it where it is.

80 00:12:48.000 00:12:59.860 Amber Lin: Okay? I guess another question we have is, how are we going to do it when we had different departments? So there’s mechanical. And that does that. Does rag help with that, too.

81 00:12:59.860 00:13:17.829 Mustafa Raja: Yeah, what we can do is we can add each department as as a field in metadata. So we can say, we can have an agent prior to our current agent. To determine which department is on asking the question.

82 00:13:18.474 00:13:25.419 Mustafa Raja: And then filter. Only the content. That’s in the rag based on the department.

83 00:13:29.170 00:13:34.640 Amber Lin: Oh, oh, wow! Are we still is this rag is this.

84 00:13:34.640 00:13:35.170 Mustafa Raja: Yeah, this.

85 00:13:35.170 00:13:40.180 Amber Lin: That’s all we need. Wow, all right.

86 00:13:40.570 00:13:42.420 Amber Lin: Yeah, I’ve been very nice.

87 00:13:43.337 00:13:49.599 Amber Lin: Let me think. What was the other one? How to schedule. An estimate

88 00:13:50.890 00:13:53.020 Amber Lin: did not spell that correctly.

89 00:14:06.850 00:14:15.190 Amber Lin: Okay, yeah, this is really, this is so much better. And let me check one last thing, estimate same day.

90 00:14:27.530 00:14:31.100 Mustafa Raja: We can also test for the animal in the toilet thing.

91 00:14:31.752 00:14:37.670 Amber Lin: It’s so funny. That’s the one thing I remember, because it was so funny.

92 00:14:45.070 00:14:51.589 Amber Lin: Yeah, I think there was a there was a URL in this and didn’t include it. But it was. It’s already really nice

93 00:14:51.900 00:14:56.305 Amber Lin: cancellations. Okay? Moving that.

94 00:14:57.520 00:14:59.880 Amber Lin: Okay.

95 00:15:00.510 00:15:10.399 Amber Lin: hey? They’re pretty nice. This is already a much better than what we had before. So I I would say, it’s a. It was a improvement. Because before we did not have these scripts.

96 00:15:14.190 00:15:17.280 Amber Lin: Okay.

97 00:15:25.150 00:15:46.279 Amber Lin: okay, I don’t know if this is what they put. I don’t think we have that specific documentation. Yeah, I think it’s just if it probably right now, whenever it’s not in the Central Doc, and he still kind of gives an answer like it’s still it still tries, and then it comes out with very generic things, but overall, whenever it is in the central doc like this, it it does really well.

98 00:15:48.530 00:15:49.639 Casie Aviles: Yeah, I see, I see.

99 00:15:50.300 00:16:00.089 Amber Lin: So I guess the only thing would be like if it’s not, just say it’s just say it’s not or suggest, like, maybe go to that section and check if something similar. But that’s good

100 00:16:02.050 00:16:09.434 Amber Lin: I’m that’s really great. Honestly, I didn’t. I didn’t even expect this to become so much better overnight. So thank you guys for doing that.

101 00:16:12.370 00:16:13.510 Amber Lin: Oh.

102 00:16:13.620 00:16:21.139 Amber Lin: I looked at this is what Chat Gpt gave. I think, what what you you guys said. We’re gonna do this

103 00:16:21.400 00:16:38.180 Amber Lin: or hierarch. I I just want to note down what we agreed on. The approach is, and we can have different tests or prototypes that we can test out again. Each could be like a 3 3 point to 5 point like test. And then we can see what works best.

104 00:16:39.030 00:16:41.449 Amber Lin: I just want to note down what you guys think.

105 00:16:42.330 00:16:50.440 Mustafa Raja: Yeah, Casey, what do you think about adding the departments as metadata my finger on it.

106 00:16:51.220 00:16:52.636 Casie Aviles: Yeah, we could do that.

107 00:16:53.190 00:16:56.120 Casie Aviles: let me just, I guess, to clarify like.

108 00:16:56.410 00:16:58.510 Casie Aviles: how how does it look like

109 00:16:59.028 00:17:00.580 Casie Aviles: like? Where? Where do you mean.

110 00:17:01.290 00:17:08.929 Amber Lin: Yeah. So they just shared with me. Some of their.

111 00:17:09.640 00:17:11.650 Amber Lin: They also have a lot of stuff.

112 00:17:11.770 00:17:19.580 Amber Lin: So I saw they shared 2 folders. You also have access as our shared email, and then each of them has a lot of documents.

113 00:17:19.789 00:17:20.429 Casie Aviles: And you click.

114 00:17:20.430 00:17:35.760 Amber Lin: And there’s more documents. So we cannot do contacts. And then there’s new hire training, which also has some stuff. But this is less important, but they have a lot, so I guess we can. We’ll have to put everything

115 00:17:36.360 00:17:42.870 Amber Lin: we can put everything in the central dock, or we can directly put all of these into super base

116 00:17:44.510 00:17:46.660 Amber Lin: what do you guys think.

117 00:17:48.735 00:17:54.779 Mustafa Raja: I just want to know. The both departments are going to be talking to the same bot.

118 00:17:55.313 00:17:57.079 Amber Lin: Yes, that’s what they want.

119 00:17:57.080 00:18:04.569 Mustafa Raja: Yeah. So so Andy needs to figure out the 1st thing Andy Andy needs to figure out is which department is asking the question right?

120 00:18:06.340 00:18:16.790 Mustafa Raja: Yeah, so that that’s going to be our main problem the the next thing. Once, Andy figures out which department it is, I feel

121 00:18:17.540 00:18:19.000 Mustafa Raja: I feel we

122 00:18:19.210 00:18:27.879 Mustafa Raja: putting a metadata filter on department in the rag would be the way to go. What do you.

123 00:18:27.880 00:18:32.469 Casie Aviles: Yeah, yeah, we could do that. So so like, we have one central.

124 00:18:33.070 00:18:33.580 Mustafa Raja: Yeah.

125 00:18:33.580 00:18:34.879 Casie Aviles: Is that what you mean?

126 00:18:35.180 00:18:37.797 Mustafa Raja: Yeah, or we can, we can have

127 00:18:38.100 00:18:40.869 Casie Aviles: Multiple tables, multiple multiple drags.

128 00:18:42.080 00:18:45.999 Mustafa Raja: But main thing is figuring out which department is asking the question.

129 00:18:46.310 00:18:46.890 Casie Aviles: Yeah, as long.

130 00:18:46.890 00:18:48.429 Amber Lin: How would we can do that?

131 00:18:51.730 00:18:54.690 Mustafa Raja: Hmm any ideas, Casey.

132 00:18:55.960 00:18:57.489 Casie Aviles: Yeah, I’m also thinking.

133 00:18:58.816 00:19:00.799 Amber Lin: I can. Yeah, go ahead.

134 00:19:01.140 00:19:07.369 Casie Aviles: Yeah, I mean, the 1st thing I I could think of is based on the questions like

135 00:19:09.070 00:19:11.260 Casie Aviles: we could have the bot. You know.

136 00:19:12.710 00:19:19.620 Casie Aviles: we’re gonna give instructions so that it determines from which department. So if, of of course, if it’s like, you know the usual

137 00:19:20.873 00:19:27.429 Casie Aviles: best questions, it’s gonna identify it. As you know, it’s gonna use the pest central dock.

138 00:19:28.620 00:19:30.730 Casie Aviles: but that’s mostly just prompting.

139 00:19:31.556 00:19:42.990 Amber Lin: I know, I remember, back then also brought up that cause. We have we have the emails of Whoever is in which department.

140 00:19:44.740 00:19:47.850 Amber Lin: So there’s people who’s just in

141 00:19:48.120 00:19:57.020 Amber Lin: test. And there’s people who’s just a mechanical. And then there’s a few people who’s like in pets and

142 00:19:57.880 00:20:07.770 Amber Lin: overflows for mechanical. So they answer both questions. So like, if we use those emails, then most cases.

143 00:20:07.910 00:20:10.340 Amber Lin: pest people would just ask pest questions.

144 00:20:10.790 00:20:11.330 Mustafa Raja: Yeah.

145 00:20:11.330 00:20:18.059 Casie Aviles: Yeah, I think using the emails can add some determinism that or like, you know, it’s easier to

146 00:20:18.550 00:20:21.430 Casie Aviles: figure out who is, you know, which department.

147 00:20:21.800 00:20:22.440 Amber Lin: Hmm!

148 00:20:22.930 00:20:25.520 Amber Lin: Or maybe we can have something that’s just

149 00:20:25.750 00:20:34.599 Amber Lin: like at the start of questions. Maybe they click a button to say Pest! And then, whatever’s pest? Do you think that works.

150 00:20:38.730 00:20:43.830 Casie Aviles: Actually, I mean, yeah, I think that’s also possible. Okay.

151 00:20:44.100 00:20:49.829 Casie Aviles: from for. So what they want is they just want one interface, right? Like just one Andy, to talk to.

152 00:20:50.313 00:20:53.139 Amber Lin: Maybe behind the scenes we could have multiple.

153 00:20:55.640 00:20:57.800 Casie Aviles: We can route to multiple

154 00:21:00.290 00:21:04.380 Casie Aviles: Llms that are for per like per department.

155 00:21:04.960 00:21:16.290 Casie Aviles: That could be something. But I guess I’m just thinking if that might introduce more stuff to.

156 00:21:16.710 00:21:17.050 Amber Lin: I see.

157 00:21:17.050 00:21:21.040 Casie Aviles: See you, debug, or like more more time, I guess.

158 00:21:21.040 00:21:29.409 Amber Lin: That’s true, right? Right here, it says, using all the 1st 3, we said. And it says also session, memory.

159 00:21:30.130 00:21:41.330 Amber Lin: so, or initial department tagging and then department.

160 00:21:41.580 00:21:42.900 Amber Lin: Wait.

161 00:21:45.360 00:21:47.820 Amber Lin: No, I don’t think that makes sense.

162 00:21:48.270 00:21:48.820 Casie Aviles: Or like.

163 00:21:48.820 00:21:49.839 Amber Lin: Guess when they.

164 00:21:50.250 00:21:51.880 Mustafa Raja: Yeah, it’s like a.

165 00:21:51.880 00:21:53.630 Amber Lin: User preference. I guess.

166 00:21:54.090 00:21:58.529 Mustafa Raja: Yeah, yeah, what do we think? What? What do we think about? What if we do?

167 00:21:59.674 00:22:07.229 Mustafa Raja: The initial routing agent, if if it cannot figure out which department is being asked the question.

168 00:22:08.470 00:22:12.836 Mustafa Raja: We can actually really send it to the rag either way, because

169 00:22:13.530 00:22:18.422 Mustafa Raja: it’s going to answer with the with with the embeddings right.

170 00:22:18.830 00:22:19.690 Casie Aviles: Don’t match.

171 00:22:19.690 00:22:25.830 Mustafa Raja: Yeah, it’s going to match and whichever department has that kind of content

172 00:22:26.110 00:22:31.350 Mustafa Raja: it will fetch from there, or we can add a human in the loop thing.

173 00:22:31.590 00:22:36.940 Mustafa Raja: If the if the routing thing does not

174 00:22:37.679 00:22:46.899 Mustafa Raja: is not able to identify which department is being asked the question, we can add a human in the loop. Ask them, which department are they from?

175 00:22:47.600 00:22:48.889 Amber Lin: Yeah. I think.

176 00:22:48.890 00:22:51.950 Mustafa Raja: Email email thing would narrow, narrow down a lot of.

177 00:22:52.420 00:22:54.360 Mustafa Raja: A lot of them is, some of them.

178 00:22:54.360 00:22:54.700 Amber Lin: Yeah.

179 00:22:54.700 00:22:58.380 Mustafa Raja: Who are overlapping this would be further my field.

180 00:22:58.380 00:23:23.810 Amber Lin: I see, I see, I think this is great. So we’ll 1st do a email, and then second, if we’re not sure we’ll try to detect the intent. And then, if we’re really not sure and then we’ll ask like, we’ll we’ll send a message. Say, are you asking about pest control? Mechanical. And then also we can use previous session memories to see like what this person was asked. Like what department they’re under.

181 00:23:24.120 00:23:24.740 Mustafa Raja: Yeah.

182 00:23:25.780 00:23:29.949 Amber Lin: Okay, that’s good. And then.

183 00:23:30.420 00:23:31.979 Mustafa Raja: How do you feel about this Kissy?

184 00:23:33.400 00:23:42.040 Casie Aviles: Yeah, this is fine. Yeah, yeah. Yeah. Or like, do they have? What do you call this?

185 00:23:42.570 00:23:44.130 Casie Aviles: Do they have like a list of

186 00:23:45.240 00:23:49.179 Casie Aviles: people who they expect to use? Or is it constantly changing.

187 00:23:49.481 00:23:51.890 Amber Lin: They should have a set list of people.

188 00:23:53.300 00:23:55.819 Casie Aviles: And yeah, and from which department they’re from right?

189 00:23:55.820 00:23:57.939 Amber Lin: Yeah, each. Department, yeah.

190 00:23:58.830 00:23:59.260 Casie Aviles: Maybe.

191 00:23:59.260 00:23:59.920 Mustafa Raja: That.

192 00:24:00.220 00:24:04.010 Casie Aviles: Do you think we could also have a list of that, and maybe have it in.

193 00:24:04.640 00:24:05.810 Amber Lin: Yeah. Totally.

194 00:24:06.370 00:24:06.960 Casie Aviles: Yeah.

195 00:24:08.300 00:24:10.019 Amber Lin: And overflows.

196 00:24:16.880 00:24:18.000 Amber Lin: Okay,

197 00:24:28.490 00:24:30.110 Amber Lin: yeah, we could do

198 00:24:30.950 00:24:48.869 Amber Lin: can do like a simple routing logic. We can test it because we have all of the all the mechanical docs. Now, probably next week we’ll put we’ll just copy and paste it into one Central Doc, or I don’t know how we’re gonna put it in and then get it super base. And we can do a quick

199 00:24:48.990 00:24:50.570 Amber Lin: can do a quick test

200 00:24:50.830 00:24:57.520 Amber Lin: like it doesn’t even have to be all the documents. We can just test a quick prototype if, like our routing logic works.

201 00:24:57.520 00:25:04.130 Casie Aviles: Yes, I think I I like the how the Central Dock is using headings now, because.

202 00:25:04.130 00:25:04.710 Amber Lin: Hmm.

203 00:25:05.020 00:25:05.850 Casie Aviles: With the

204 00:25:06.050 00:25:13.480 Casie Aviles: yeah, with the processing, with the script that I have for process for chunking the central Doc, it was easily able to.

205 00:25:14.727 00:25:22.460 Amber Lin: You know, they split it into sections. So I think that was a good dutch. Yeah, okay.

206 00:25:23.020 00:25:28.089 Amber Lin: do. We still need to add the departments, and as metadata in the.

207 00:25:29.550 00:25:35.219 Mustafa Raja: Yeah, if we if we are not making multiple drags.

208 00:25:35.450 00:25:40.760 Mustafa Raja: and we want to work in the same drag, then we will have to have metadata.

209 00:25:40.760 00:25:45.690 Amber Lin: I see what are the benefits and downsides of having it separate and.

210 00:25:45.690 00:25:50.270 Mustafa Raja: Yeah, the only downside. Yeah. The only downsides that I can think of.

211 00:25:51.173 00:25:56.070 Mustafa Raja: If more number of rows could slow down the response time.

212 00:25:56.170 00:25:58.640 Amber Lin: But we will have to see if.

213 00:25:59.290 00:26:04.559 Mustafa Raja: If that would be the case, I’m not sure. I’m not sure if number of rows would affect

214 00:26:05.060 00:26:06.340 Mustafa Raja: it at all.

215 00:26:07.226 00:26:08.399 Amber Lin: I see.

216 00:26:09.020 00:26:11.190 Amber Lin: So it says.

217 00:26:12.250 00:26:24.240 Amber Lin: well, it says, cause good talking discipline. Will, it slow down rag, if we have many rows, question mark.

218 00:26:26.720 00:26:29.890 Amber Lin: Here, this step? 2.

219 00:26:42.040 00:26:44.519 Amber Lin: Hey? What does it? What does it say?

220 00:26:45.640 00:26:50.189 Amber Lin: Vector, if we’re using option? One, which is a single

221 00:26:50.650 00:26:57.510 Amber Lin: one single one single rag vector to be depends on how selective.

222 00:26:57.510 00:26:58.210 Mustafa Raja: So these.

223 00:26:59.506 00:27:11.350 Amber Lin: Okay, I we are using super base. Let me see if it changes anything. Of

224 00:27:12.298 00:27:14.599 Amber Lin: this is not what I asked for.

225 00:27:16.430 00:27:18.039 Casie Aviles: Spit out, code suddenly.

226 00:27:19.630 00:27:27.500 Amber Lin: Anything question mark. So I guess it says it doesn’t fill. It doesn’t change.

227 00:27:27.960 00:27:33.010 Amber Lin: It’s just how big our filter is.

228 00:27:34.020 00:27:34.660 Mustafa Raja: Yeah.

229 00:27:35.620 00:27:41.100 Mustafa Raja: So I saw that it’s only 150 around 150 rows for the central dog.

230 00:27:41.300 00:27:44.650 Amber Lin: Yes, yeah. I don’t know.

231 00:27:44.650 00:27:48.079 Mustafa Raja: Fine, how many, how, how? How big is the

232 00:27:48.710 00:27:51.919 Mustafa Raja: what’s it called knowledge base for the mechanical department?

233 00:27:51.920 00:27:58.450 Amber Lin: Probably similar. Like definitely under 500, it will probably be like 200 ish.

234 00:27:59.030 00:28:07.799 Mustafa Raja: Yeah, then, yeah, then, it’s going to be 500 at most. The number of floors which will be filtered down by the department names.

235 00:28:07.950 00:28:09.830 Amber Lin: So I feel pretty good about this.

236 00:28:11.490 00:28:16.049 Amber Lin: Hmm, okay, yeah. It says here.

237 00:28:16.280 00:28:29.050 Amber Lin: if it’s 500 K docs with index filters. No, we’ll slow down if it’s across 100 K documents per department. Which I we’re not getting there anytime. Soon.

238 00:28:29.050 00:28:32.150 Mustafa Raja: Yeah, okay, that sounds good.

239 00:28:32.150 00:28:36.400 Mustafa Raja: And we we would. Then, if the if it if it gets to that point, we could

240 00:28:36.620 00:28:38.359 Mustafa Raja: split up the ranks.

241 00:28:38.850 00:28:48.869 Amber Lin: Yeah, okay, I guess after that, are we setting? We? I just want to list out a few things we need to set up. We need to set up like the feedback

242 00:28:49.210 00:28:50.050 Amber Lin: loop.

243 00:28:51.521 00:29:05.650 Amber Lin: So that the Csrs. I think it’s the same thing. But we need to route the feedback loop route feedback loop for mechanical cause. I don’t know if we want everything to be inside the same feedback sheet.

244 00:29:07.232 00:29:11.160 Amber Lin: I don’t know like that’s something optional we can do.

245 00:29:11.630 00:29:17.680 Amber Lin: Was there any other thing that we’re doing? Are we doing? Do you? Do we want to do evals for

246 00:29:17.850 00:29:22.129 Amber Lin: mechanical? Not that our current evils really does anything.

247 00:29:24.514 00:29:30.659 Casie Aviles: Yeah, but maybe we still should.

248 00:29:30.780 00:29:33.510 Casie Aviles: But just so, yeah. But I think

249 00:29:34.260 00:29:36.340 Casie Aviles: the the what do you call this?

250 00:29:36.770 00:29:39.259 Casie Aviles: We still need to address the route which is

251 00:29:40.260 00:29:42.749 Casie Aviles: our evals need to be improved. I guess.

252 00:29:43.980 00:29:46.410 Casie Aviles: Okay, we’re yeah.

253 00:29:48.850 00:29:56.746 Amber Lin: Yeah, no, there, the evil numbers make notes. It doesn’t matter right now. It’s always like 8 or 9 or 10. It’s just that’s it.

254 00:29:58.110 00:29:59.165 Amber Lin: Okay,

255 00:30:00.500 00:30:03.140 Amber Lin: So we’ll go with option one. I’ll

256 00:30:03.630 00:30:13.230 Amber Lin: make a roadmap. I’ll talk Witham just to confirm. He’s happy with this, and then I guess next week. We can just do a quick test and prototypes.

257 00:30:13.600 00:30:14.290 Casie Aviles: Okay.

258 00:30:14.810 00:30:17.099 Amber Lin: Okay, thank you. All.

259 00:30:17.240 00:30:18.080 Mustafa Raja: Thank you.

260 00:30:18.080 00:30:25.859 Casie Aviles: Wait sorry amber just to clarify again. Which are there any tickets that you want me to close of within.

261 00:30:27.870 00:30:32.850 Amber Lin: I guess not. Within today. It’s just if you can start on this one.

262 00:30:33.440 00:30:34.290 Casie Aviles: Service area.

263 00:30:34.290 00:30:45.410 Amber Lin: Service. Yeah, it will be probably will be the same as inspectors. It’s the I copied over the zip codes from the service area sheet. So it’s actually the same zip codes.

264 00:30:45.780 00:30:50.590 Casie Aviles: I see. Yeah, it’s in the master spreadsheet, right? I mean the spreadsheet hub as well.

265 00:30:51.213 00:30:59.900 Amber Lin: No, maybe I think I I think I copied it over. So yes, it’s over.

266 00:31:02.060 00:31:04.139 Amber Lin: Here it’s over here.

267 00:31:05.410 00:31:08.260 Amber Lin: There’s ABC comfree.

268 00:31:08.800 00:31:16.810 Amber Lin: Yeah, I’ll probably copy. Okay, chem free is pretty small. So it’s just ABC, and then.

269 00:31:16.810 00:31:18.139 Casie Aviles: Okay, I’m free. Okay.

270 00:31:18.140 00:31:18.680 Amber Lin: Yeah.

271 00:31:21.930 00:31:26.250 Amber Lin: oh, what is this? Okay, just these 2, these 2.

272 00:31:26.250 00:31:26.920 Casie Aviles: All right.

273 00:31:26.920 00:31:29.120 Amber Lin: Copy. ABC, copy of service.

274 00:31:29.330 00:31:30.500 Amber Lin: Yeah, all.

275 00:31:31.040 00:31:38.390 Amber Lin: I’ll say that it’s and spreadsheet hub.

276 00:31:39.350 00:31:40.140 Amber Lin: Okay.

277 00:31:44.860 00:31:51.990 Amber Lin: yeah. Just a quick test to see if it works, and then, once we test it, I can give more feedback.

278 00:31:53.240 00:31:53.860 Casie Aviles: Okay.

279 00:31:54.450 00:31:55.290 Amber Lin: All right.

280 00:31:56.720 00:31:58.169 Casie Aviles: Yeah, that should be it for me.

281 00:31:58.730 00:31:59.400 Amber Lin: Okay.

282 00:32:01.180 00:32:02.620 Amber Lin: Thank you guys.

283 00:32:02.990 00:32:03.770 Mustafa Raja: Thank you.

284 00:32:03.770 00:32:04.179 Casie Aviles: Thank you.

285 00:32:04.830 00:32:06.180 Amber Lin: Alright, bye.