Meeting Title: AI Team Standup Date: 2025-04-28 Meeting participants: Amber Lin, Miguel De Veyra, Casie Aviles, Awaish Kumar
WEBVTT
1 00:01:08.380 ⇒ 00:01:12.210 Amber Lin: Hi, there, let me share my screen.
2 00:01:14.362 ⇒ 00:01:19.060 Amber Lin: Let’s see. So being my issues.
3 00:01:20.630 ⇒ 00:01:28.759 Amber Lin: So the main request from Utah, and what we want to focus on is, when can we have agents out?
4 00:01:29.860 ⇒ 00:01:31.530 Amber Lin: So what do you guys think.
5 00:01:38.020 ⇒ 00:01:41.220 Miguel de Veyra: Okay, think we should be able to.
6 00:01:43.170 ⇒ 00:01:46.049 Miguel de Veyra: So I know me and Casey worked on an agent
7 00:01:46.160 ⇒ 00:01:49.070 Miguel de Veyra: getting an agent out last week. So.
8 00:01:50.180 ⇒ 00:01:51.490 Amber Lin: Which one was that?
9 00:01:52.540 ⇒ 00:01:54.180 Amber Lin: Oh, the law firm! One.
10 00:01:56.260 ⇒ 00:02:02.220 Miguel de Veyra: Yeah, yeah, I think I think we list down what agency he wants first.st So we can prioritize that.
11 00:02:02.220 ⇒ 00:02:06.549 Amber Lin: Yeah, interesting, is it like.
12 00:02:06.550 ⇒ 00:02:08.470 Miguel de Veyra: The client. Agency avi.
13 00:02:08.479 ⇒ 00:02:10.969 Amber Lin: I mean agents we have.
14 00:02:11.579 ⇒ 00:02:20.129 Amber Lin: She is done. We have jar, we kind of eaten all parts.
15 00:02:20.909 ⇒ 00:02:23.759 Amber Lin: Our staff.
16 00:02:28.779 ⇒ 00:02:30.439 Amber Lin: What other projects?
17 00:02:31.969 ⇒ 00:02:39.009 Amber Lin: Gosh, that’s it. Said it.
18 00:02:39.010 ⇒ 00:02:42.100 Casie Aviles: I think they’re. They also have matter more right.
19 00:02:42.130 ⇒ 00:02:43.949 Amber Lin: Oh, yeah, right?
20 00:02:47.060 ⇒ 00:02:49.570 Amber Lin: So there’s the different agents.
21 00:02:51.570 ⇒ 00:02:57.220 Amber Lin: These I’m on. So I can assess, if answers are correct for these
22 00:02:57.350 ⇒ 00:03:00.330 Amber Lin: these we will have to ask Robert.
23 00:03:00.880 ⇒ 00:03:02.610 Amber Lin: So these are.
24 00:03:02.610 ⇒ 00:03:09.080 Casie Aviles: I I think, Javi, I’m not sure but this I think Javi is not active right now. Right.
25 00:03:12.620 ⇒ 00:03:17.599 Amber Lin: Not active right now. Yes, but they’re planning for a renewal.
26 00:03:17.940 ⇒ 00:03:21.590 Amber Lin: so I could say, Move it down.
27 00:03:22.710 ⇒ 00:03:29.930 Casie Aviles: Because I just, you know I guess my concern is just to work on what the agent that we could get the most feedback.
28 00:03:30.080 ⇒ 00:03:30.440 Amber Lin: Yeah.
29 00:03:30.440 ⇒ 00:03:37.870 Casie Aviles: Like sooner. So yeah, unfortunately, knowing that job is not, yeah, okay, yeah.
30 00:03:38.340 ⇒ 00:03:51.079 Amber Lin: Eden’s very complex, I would say might not be the greatest 1st target. Also, feedback from Robert might be a bit slow, because he’s really busy. This is very small.
31 00:03:51.190 ⇒ 00:03:55.660 Amber Lin: Oh, very small. Project just started.
32 00:03:56.725 ⇒ 00:04:04.020 Amber Lin: Urban stamps. This is mostly demote and hotel
33 00:04:06.269 ⇒ 00:04:15.139 Amber Lin: so you probably need to get feedback from them direct, sometimes from them, because a lot of the details I’m not that sure of.
34 00:04:15.727 ⇒ 00:04:22.469 Amber Lin: This is amber loop, so you can ask me on the feedback. This is Bobby. You have to ask Robert
35 00:04:22.630 ⇒ 00:04:23.929 Amber Lin: for feedback.
36 00:04:25.450 ⇒ 00:04:26.070 Casie Aviles: Okay.
37 00:04:27.200 ⇒ 00:04:33.489 Miguel de Veyra: Yeah, I have a question, though, Casey. And this once, which data do we have like from this clients?
38 00:04:36.756 ⇒ 00:04:37.823 Casie Aviles: For these clients.
39 00:04:38.180 ⇒ 00:04:38.730 Miguel de Veyra: Yeah, yeah.
40 00:04:38.730 ⇒ 00:04:41.929 Casie Aviles: But I do know that we have the data for Javi.
41 00:04:43.270 ⇒ 00:04:47.439 Casie Aviles: I think I need to recheck all of the data that we have, because
42 00:04:48.491 ⇒ 00:04:55.728 Casie Aviles: I don’t think all of it is complete. But I do know that we have the Zoom Meeting data complete.
43 00:04:56.420 ⇒ 00:04:57.900 Casie Aviles: the slack data.
44 00:04:58.650 ⇒ 00:05:00.959 Casie Aviles: So all we need is to just filter.
45 00:05:02.550 ⇒ 00:05:05.970 Miguel de Veyra: Wait. That’s for all the meetings for all these clients. Right?
46 00:05:06.240 ⇒ 00:05:09.170 Miguel de Veyra: We have the data for that in super base.
47 00:05:12.420 ⇒ 00:05:17.320 Casie Aviles: Oh, in super base. No, they’re not yet on super base, except for zoom. So Zoom has
48 00:05:18.250 ⇒ 00:05:20.319 Casie Aviles: our stuff in super base.
49 00:05:21.930 ⇒ 00:05:28.159 Miguel de Veyra: This is sorry for zoom in super base, that’s all meetings already, or just a specific amount of meetings.
50 00:05:31.460 ⇒ 00:05:35.630 Casie Aviles: Just the specific amount, because there’s we have, like
51 00:05:35.870 ⇒ 00:05:41.629 Casie Aviles: all other meetings in S 3, right? And that would mean we have to vectorize all of those.
52 00:05:42.100 ⇒ 00:05:42.950 Amber Lin: Hmm.
53 00:05:43.990 ⇒ 00:05:50.939 Miguel de Veyra: Oh, yeah, but we didn’t. We discuss like last week that it’s only the ex. It’s only the new ones that will process for now or.
54 00:05:51.250 ⇒ 00:05:59.489 Casie Aviles: Yes, it’s only no, no, that’s correct. So for each Zoom Meeting that comes like, for example, the one earlier.
55 00:06:00.130 ⇒ 00:06:06.130 Casie Aviles: It will get vectorized. And then it will be added to super base. But they’re just, you know, they’re older
56 00:06:06.380 ⇒ 00:06:12.230 Casie Aviles: transcripts that we’ll have to still get
57 00:06:12.730 ⇒ 00:06:17.220 Casie Aviles: right because there there are videos dating back to 2023.
58 00:06:22.440 ⇒ 00:06:25.469 Amber Lin: Not so. The old videos are not in super base yet.
59 00:06:27.235 ⇒ 00:06:27.740 Casie Aviles: Yes.
60 00:06:28.770 ⇒ 00:06:38.049 Amber Lin: I see. So what what type of Zoom Meetings are in super base? Is it client specific, like? Do we only have Javi or ABC.
61 00:06:38.830 ⇒ 00:06:39.649 Amber Lin: But it’s been.
62 00:06:40.007 ⇒ 00:06:47.510 Casie Aviles: All meetings that regardless of the client, you know all the recordings that we have coming in that is new.
63 00:06:47.780 ⇒ 00:06:48.660 Amber Lin: Oh, yeah.
64 00:06:49.030 ⇒ 00:06:49.950 Casie Aviles: Characterize.
65 00:06:50.817 ⇒ 00:06:53.490 Miguel de Veyra: Sorry, Casey, just sorry to ask you again, but
66 00:06:53.990 ⇒ 00:06:56.780 Miguel de Veyra: in the metadata did we add the client, name or no?
67 00:06:57.430 ⇒ 00:07:00.340 Casie Aviles: Yeah, it’s already there. I you mean the client’s name.
68 00:07:00.848 ⇒ 00:07:06.770 Casie Aviles: Oh, no, no client name, no sorry client name. No, but we do have the meeting name.
69 00:07:07.420 ⇒ 00:07:14.920 Casie Aviles: And yeah, we do have the meeting name and the meeting date, and then the meeting category.
70 00:07:17.540 ⇒ 00:07:23.990 Miguel de Veyra: Okay, okay, yeah, I think. Cause I think for this, we need to filter a lot. Now, via metadata.
71 00:07:25.640 ⇒ 00:07:26.170 Amber Lin: Oh!
72 00:07:27.130 ⇒ 00:07:27.980 Miguel de Veyra: Okay. Yeah.
73 00:07:28.440 ⇒ 00:07:30.550 Amber Lin: Sounds good can comment that.
74 00:07:30.830 ⇒ 00:07:34.090 Amber Lin: So that’s for zoom. What about for slack.
75 00:07:38.478 ⇒ 00:07:42.950 Casie Aviles: This one we have in factorized. So that means it’s still not on super base.
76 00:07:44.147 ⇒ 00:07:49.439 Casie Aviles: We’re thinking of, like, what’s the best way, we could do this, that is.
77 00:07:51.100 ⇒ 00:07:54.950 Casie Aviles: it’s going to be the most helpful for the AI because
78 00:07:55.820 ⇒ 00:07:57.680 Casie Aviles: I guess the problem is that
79 00:07:58.467 ⇒ 00:08:05.440 Casie Aviles: so for each data source. There’s like a different structure for each. So Zoom has a different structure.
80 00:08:05.740 ⇒ 00:08:07.730 Casie Aviles: Slack has a different structure.
81 00:08:08.180 ⇒ 00:08:14.000 Casie Aviles: So what’s like the best way that we could put it inside super base
82 00:08:14.740 ⇒ 00:08:17.010 Casie Aviles: and make it accessible to the AI.
83 00:08:17.950 ⇒ 00:08:19.236 Amber Lin: I see
84 00:08:19.880 ⇒ 00:08:23.949 Miguel de Veyra: Thinking for that. Casey can remember we have this, the
85 00:08:24.480 ⇒ 00:08:28.719 Miguel de Veyra: what do you call it? The basically how many messages were sent? Tracker.
86 00:08:30.240 ⇒ 00:08:31.560 Casie Aviles: Yeah, we do have that.
87 00:08:31.560 ⇒ 00:08:32.730 Miguel de Veyra: Yeah, we do have that.
88 00:08:32.940 ⇒ 00:08:39.810 Miguel de Veyra: Can we do that? And then basically, not daily, of course, but like, or maybe daily, just
89 00:08:40.260 ⇒ 00:08:45.089 Miguel de Veyra: same path. Because I’m I’m I’m assuming. If we can count it, we have the context right?
90 00:08:45.815 ⇒ 00:08:54.960 Miguel de Veyra: We, we have the messages and then basically do the same in zoom where it’s in the end workflow that does the vectorizing, and then just adds it into
91 00:08:55.120 ⇒ 00:08:59.219 Miguel de Veyra: super into super base. Right? I think that’s the best way we can do that.
92 00:09:00.800 ⇒ 00:09:05.620 Casie Aviles: Okay. So you mean, like, for incoming messages or daily messages, I mean.
93 00:09:06.180 ⇒ 00:09:11.699 Miguel de Veyra: Yeah. Yeah. Or maybe, if cause, I don’t think there’s like really a lot of conversation in across the channels.
94 00:09:12.270 ⇒ 00:09:19.659 Miguel de Veyra: So I’m thinking we we can filter it like a I mean, yeah, daily is good. But
95 00:09:19.870 ⇒ 00:09:25.569 Miguel de Veyra: I’m thinking, just getting like the past few months how we can do it.
96 00:09:25.740 ⇒ 00:09:27.180 Miguel de Veyra: But yeah, something like that.
97 00:09:29.780 ⇒ 00:09:31.180 Casie Aviles: Oh, okay.
98 00:09:32.290 ⇒ 00:09:37.990 Miguel de Veyra: I think that’s a bit more plausible than having to depend on the data. And next 3, right?
99 00:09:39.520 ⇒ 00:09:45.219 Miguel de Veyra: Because we so we if we depend on data from Sd, we still need to collect it. I mean to process it.
100 00:09:45.580 ⇒ 00:09:49.030 Miguel de Veyra: And then, of course, that’s gonna push back our agent time again.
101 00:09:52.780 ⇒ 00:09:58.800 Amber Lin: Is there some? Is there a way? If we can just use a small batch of data and just
102 00:09:59.840 ⇒ 00:10:08.719 Amber Lin: get it into super base with some level of formatting, and then we can see the results of the agent and then be able to get feedback
103 00:10:09.415 ⇒ 00:10:10.390 Amber Lin: and be able.
104 00:10:10.940 ⇒ 00:10:18.439 Amber Lin: If this structure works. So we can just pick a a set of data for like a client.
105 00:10:19.290 ⇒ 00:10:19.800 Casie Aviles: Okay.
106 00:10:19.800 ⇒ 00:10:31.620 Amber Lin: Try a certain structure. Well, we can just try it out like we want to do. We can just figure that out, and then that will also help us say that we can deliver a agent.
107 00:10:32.530 ⇒ 00:10:33.559 Amber Lin: so we don’t have to get.
108 00:10:33.560 ⇒ 00:10:33.970 Casie Aviles: Yeah.
109 00:10:33.970 ⇒ 00:10:35.350 Amber Lin: That’s okay.
110 00:10:36.340 ⇒ 00:10:41.949 Casie Aviles: That makes sense, I think. Yeah, we could do that definitely some focusing on a smaller batch. For now.
111 00:10:43.660 ⇒ 00:10:45.969 Amber Lin: It’s a small badge.
112 00:10:46.560 ⇒ 00:10:48.040 Miguel de Veyra: Can we take matter more?
113 00:10:48.470 ⇒ 00:10:50.850 Miguel de Veyra: Sure? Yeah, yeah, that’s awesome.
114 00:10:52.270 ⇒ 00:10:54.210 Miguel de Veyra: So there’s like, not a lot of backlog.
115 00:10:54.210 ⇒ 00:10:59.180 Amber Lin: Yeah, this really is very, very minimal messages.
116 00:11:07.120 ⇒ 00:11:08.210 Amber Lin: vector.
117 00:11:12.110 ⇒ 00:11:14.550 Amber Lin: let me see, okay.
118 00:11:14.810 ⇒ 00:11:15.480 Casie Aviles: Yes.
119 00:11:16.150 ⇒ 00:11:23.500 Amber Lin: Awesome. So if we do that, that means we can have matter more agent out soon, right? At least a test version.
120 00:11:24.980 ⇒ 00:11:25.630 Casie Aviles: Okay.
121 00:11:25.770 ⇒ 00:11:26.610 Casie Aviles: Yes.
122 00:11:26.960 ⇒ 00:11:29.010 Amber Lin: Okay, awesome.
123 00:11:29.960 ⇒ 00:11:36.859 Amber Lin: And what about linear? And Github?
124 00:11:40.630 ⇒ 00:11:44.550 Amber Lin: Do we have those information? How’s the progress on that.
125 00:11:46.780 ⇒ 00:11:48.429 Casie Aviles: Yeah, I believe linear is in.
126 00:11:48.430 ⇒ 00:11:50.029 Awaish Kumar: Data could be industry.
127 00:11:50.570 ⇒ 00:11:52.460 Awaish Kumar: Github is also an S. 3.
128 00:11:53.030 ⇒ 00:11:53.720 Amber Lin: Hmm.
129 00:11:56.900 ⇒ 00:11:57.780 Amber Lin: okay.
130 00:12:01.520 ⇒ 00:12:02.490 Awaish Kumar: Can you see it?
131 00:12:05.710 ⇒ 00:12:09.650 Miguel de Veyra: Yeah, yeah, we were, we were able to see the data industry.
132 00:12:10.150 ⇒ 00:12:10.800 Casie Aviles: Wait.
133 00:12:13.680 ⇒ 00:12:14.510 Casie Aviles: Okay.
134 00:12:24.760 ⇒ 00:12:33.969 Amber Lin: yeah, let’s finish that. And then we can look at the current state. So that’s an S. 3. Do we want to put the Github in super base scope.
135 00:12:34.380 ⇒ 00:12:35.949 Amber Lin: We want that to happen.
136 00:12:40.130 ⇒ 00:12:44.330 Casie Aviles: Hmm, like we we could.
137 00:12:44.750 ⇒ 00:12:50.300 Casie Aviles: you could also include it to super base. But yeah, it’s like the same
138 00:12:52.189 ⇒ 00:12:56.229 Casie Aviles: same concern that we have where we have to figure out like a structure. But.
139 00:12:58.770 ⇒ 00:13:05.559 Casie Aviles: So it’s either, do we want to prioritize 2 data sources for now? Or do you want to add another data source.
140 00:13:08.660 ⇒ 00:13:09.290 Miguel de Veyra: Thank you.
141 00:13:09.290 ⇒ 00:13:14.410 Miguel de Veyra: So I think we do. The zoom and slack. 1st get an agent deployed, and then.
142 00:13:14.410 ⇒ 00:13:14.880 Amber Lin: Good.
143 00:13:14.880 ⇒ 00:13:19.570 Miguel de Veyra: Yeah, linear, because linear isn’t made an a 10, anyways.
144 00:13:21.090 ⇒ 00:13:24.609 Miguel de Veyra: So there’s probably a way we can handle that github.
145 00:13:27.600 ⇒ 00:13:29.799 Miguel de Veyra: I don’t know. Let’s review Github later. Casey.
146 00:13:31.160 ⇒ 00:13:37.640 Amber Lin: Okay, so our goal would be, let me create a ticket.
147 00:13:42.616 ⇒ 00:13:46.480 Amber Lin: turn more client agent.
148 00:13:48.150 ⇒ 00:13:54.010 Amber Lin: Let me go here, I would say, is to do the cycle
149 00:13:54.110 ⇒ 00:13:59.810 Amber Lin: and Matt. This will allow us to one deploy agent, and 2 tests would have to.
150 00:14:00.040 ⇒ 00:14:02.010 Amber Lin: I try slack messages.
151 00:14:02.780 ⇒ 00:14:12.640 Amber Lin: and when can this be done just like a bare bones exist? It’s just for it to exist.
152 00:14:13.290 ⇒ 00:14:15.050 Amber Lin: When would this be.
153 00:14:20.310 ⇒ 00:14:25.049 Miguel de Veyra: I think, Casey, do you think we can have one by end of day or tomorrow?
154 00:14:25.640 ⇒ 00:14:26.220 Amber Lin: Wow!
155 00:14:26.220 ⇒ 00:14:32.370 Casie Aviles: Tomorrow. So we should be, yeah, because I I want to get feedback as early as possible.
156 00:14:32.370 ⇒ 00:14:33.600 Amber Lin: Hmm, okay.
157 00:14:33.850 ⇒ 00:14:36.300 Casie Aviles: And the scope is relatively small, so.
158 00:14:36.810 ⇒ 00:14:40.490 Amber Lin: Yeah. Okay, so that would be like, 3, 5.
159 00:14:40.740 ⇒ 00:14:42.160 Casie Aviles: I think that’s it.
160 00:14:42.540 ⇒ 00:14:46.530 Casie Aviles: Long as the day I think. 5, yeah, 5.
161 00:14:46.530 ⇒ 00:14:48.099 Miguel de Veyra: It has to be fine.
162 00:14:48.660 ⇒ 00:14:59.470 Amber Lin: Sounds good. Yeah, I didn’t think it was that fast. So we have. We have good. We have one agent here. I think.
163 00:14:59.730 ⇒ 00:15:01.064 Amber Lin: Oh, right there!
164 00:15:05.850 ⇒ 00:15:12.160 Amber Lin: What other agents I know there’s more agents. I just don’t really know what it is.
165 00:15:12.990 ⇒ 00:15:18.460 Miguel de Veyra: I think one of the agents that we need to. We need to fix is the.
166 00:15:19.630 ⇒ 00:15:20.139 Amber Lin: It does it.
167 00:15:20.140 ⇒ 00:15:22.599 Miguel de Veyra: Follow up. Yeah, the sales follow up.
168 00:15:24.350 ⇒ 00:15:26.300 Miguel de Veyra: I’ll work on the formatting of that.
169 00:15:29.920 ⇒ 00:15:31.999 Amber Lin: What was other ones?
170 00:15:35.960 ⇒ 00:15:38.909 Miguel de Veyra: Task. Summarizer is part of the Obby right.
171 00:15:40.270 ⇒ 00:15:41.350 Amber Lin: Yeah. That was.
172 00:15:41.350 ⇒ 00:15:41.910 Casie Aviles: Yeah, but.
173 00:15:41.910 ⇒ 00:15:45.089 Amber Lin: For every every client agent, right.
174 00:15:47.590 ⇒ 00:15:49.590 Casie Aviles: Also it has to be also part of
175 00:15:49.920 ⇒ 00:15:53.440 Casie Aviles: all agents, so the the ability to send.
176 00:15:53.740 ⇒ 00:15:55.319 Amber Lin: Oh, okay. Okay.
177 00:15:55.530 ⇒ 00:15:58.391 Miguel de Veyra: Yeah, I think that’s a feature other rather than
178 00:16:00.040 ⇒ 00:16:01.400 Casie Aviles: Then a separate agent.
179 00:16:01.800 ⇒ 00:16:02.649 Miguel de Veyra: Yeah, yeah, yeah.
180 00:16:02.650 ⇒ 00:16:04.740 Casie Aviles: So it’s more of a feature of an agent.
181 00:16:05.020 ⇒ 00:16:06.540 Miguel de Veyra: Yeah, exactly.
182 00:16:15.120 ⇒ 00:16:18.989 Miguel de Veyra: So. I think one of the things we need to add is, what.
183 00:16:19.480 ⇒ 00:16:22.390 Miguel de Veyra: for matter more, what should it have.
184 00:16:22.958 ⇒ 00:16:29.920 Amber Lin: Great, that’s a great idea. So the agents, I’ll say, we just have the sales for now, we do have that in a ticket.
185 00:16:30.590 ⇒ 00:16:33.960 Amber Lin: Where is it? Okay.
186 00:16:33.960 ⇒ 00:16:36.890 Miguel de Veyra: 1, 1, 5, 6, and to do cycle.
187 00:16:37.260 ⇒ 00:16:38.069 Amber Lin: To do this.
188 00:16:40.900 ⇒ 00:16:42.140 Amber Lin: Yeah, I’ll put it there.
189 00:16:43.147 ⇒ 00:16:51.359 Amber Lin: I think it’s high. It’s not urgent. They’re not dying because we don’t have good formatting. Let’s say, this is very high priority.
190 00:16:54.050 ⇒ 00:16:56.300 Amber Lin: Here, yeah, what’s needed.
191 00:17:01.060 ⇒ 00:17:03.689 Miguel de Veyra: Yeah, so so it has. It has to have.
192 00:17:03.690 ⇒ 00:17:04.599 Casie Aviles: A slack agent.
193 00:17:05.409 ⇒ 00:17:07.749 Miguel de Veyra: To slack and zoom.
194 00:17:11.270 ⇒ 00:17:12.969 Casie Aviles: Oh, I was thinking of the.
195 00:17:13.310 ⇒ 00:17:13.859 Amber Lin: Oh!
196 00:17:13.869 ⇒ 00:17:15.640 Casie Aviles: Yeah, bye, but the like. The.
197 00:17:15.640 ⇒ 00:17:19.230 Miguel de Veyra: Yeah, yeah, slack agent.
198 00:17:19.579 ⇒ 00:17:25.159 Miguel de Veyra: And then should it also have a task summarizer based on the meetings. For now, because we don’t have linear.
199 00:17:25.800 ⇒ 00:17:28.554 Miguel de Veyra: yeah, I think I could implement a quick one.
200 00:17:28.860 ⇒ 00:17:38.339 Miguel de Veyra: And then what else it should also contain, like the general information of the client. What who matter more is Yada Yada, what we’re doing for them.
201 00:17:39.570 ⇒ 00:17:43.840 Miguel de Veyra: because this was the this was the one who Tom brought up for Yazi. Right
202 00:17:48.810 ⇒ 00:17:50.480 Miguel de Veyra: do you think, Amber? This is enough.
203 00:17:52.486 ⇒ 00:17:56.669 Amber Lin: It. Like, very, yeah, let’s get it out. Let’s get it out.
204 00:17:57.420 ⇒ 00:17:58.479 Amber Lin: See what we need.
205 00:17:58.960 ⇒ 00:18:09.140 Casie Aviles: I I think one more. One thing that I’ll need is for the bot to be able to get data from slack it needs to be. We need to add
206 00:18:09.290 ⇒ 00:18:11.959 Casie Aviles: the brain forge, but to the.
207 00:18:11.960 ⇒ 00:18:12.330 Amber Lin: Hmm.
208 00:18:12.330 ⇒ 00:18:13.210 Casie Aviles: Channel.
209 00:18:13.890 ⇒ 00:18:17.969 Miguel de Veyra: Oh, yeah, yeah, is that something you can do? Amber.
210 00:18:18.260 ⇒ 00:18:21.379 Amber Lin: What do we need to add? Yes, I can do it.
211 00:18:22.020 ⇒ 00:18:24.950 Casie Aviles: Just the brain forge Bot, because that’s the bot.
212 00:18:25.270 ⇒ 00:18:31.999 Casie Aviles: because it has, like the credentials that I need that could help me pull the data from the channels.
213 00:18:32.250 ⇒ 00:18:35.093 Amber Lin: What is it called? Can you just
214 00:18:36.250 ⇒ 00:18:46.930 Amber Lin: I just need the name or the email to add it to a external channel. Or we can just ask Utam to add it to every single channel like, Is there automation?
215 00:18:50.150 ⇒ 00:18:50.690 Amber Lin: Yeah.
216 00:18:50.690 ⇒ 00:18:51.290 Casie Aviles: I think.
217 00:18:53.450 ⇒ 00:18:58.259 Casie Aviles: Yeah, that’s 1 of the the concerns that he raised with that approach. But
218 00:18:58.790 ⇒ 00:19:04.150 Casie Aviles: because that would mean he would add to need to have to add the bot everywhere. So
219 00:19:05.420 ⇒ 00:19:06.559 Casie Aviles: but that, yeah.
220 00:19:08.090 ⇒ 00:19:12.520 Amber Lin: Sorry did he want that to happen, or didn’t want that to happen?
221 00:19:14.018 ⇒ 00:19:20.050 Casie Aviles: Yeah, I’m not. I’m not sure yet. But yeah, I think, yeah.
222 00:19:20.660 ⇒ 00:19:22.889 Amber Lin: No worries I will.
223 00:19:23.300 ⇒ 00:19:29.240 Amber Lin: And what is the button’s name called.
224 00:19:30.310 ⇒ 00:19:32.620 Casie Aviles: Yeah, it’s just at brain forgepod.
225 00:19:35.036 ⇒ 00:19:36.269 Amber Lin: I see!
226 00:19:36.430 ⇒ 00:19:37.120 Casie Aviles: Let me!
227 00:19:37.565 ⇒ 00:19:44.079 Miguel de Veyra: Other thing we should probably do. Amber is create. I’m not sure if there’s already a matter more client project.
228 00:19:45.735 ⇒ 00:19:52.279 Amber Lin: There should be. It’s in the are you asking about the linear.
229 00:19:52.630 ⇒ 00:20:00.399 Miguel de Veyra: Yes, yes, and under aid, because I don’t think we could. No, no, no, not there. And more of like.
230 00:20:00.920 ⇒ 00:20:03.960 Miguel de Veyra: you see, add to project on the right side.
231 00:20:03.960 ⇒ 00:20:05.090 Amber Lin: Oh, yeah.
232 00:20:05.090 ⇒ 00:20:05.840 Miguel de Veyra: Yeah, yeah.
233 00:20:06.175 ⇒ 00:20:07.850 Amber Lin: I’m gonna make a oh.
234 00:20:07.850 ⇒ 00:20:10.739 Miguel de Veyra: Yeah, I don’t think we have a matter more project yet.
235 00:20:11.900 ⇒ 00:20:12.770 Amber Lin: 3.rd
236 00:20:22.730 ⇒ 00:20:24.280 Amber Lin: There we go.
237 00:20:25.880 ⇒ 00:20:31.070 Amber Lin: We can also break this down. If we want.
238 00:20:31.580 ⇒ 00:20:34.420 Amber Lin: I’ll leave that guy. I’ll leave that to you guys.
239 00:20:37.850 ⇒ 00:20:39.110 Amber Lin: Issues.
240 00:20:44.690 ⇒ 00:20:47.369 Amber Lin: Yes, and then we could get feedback
241 00:20:49.740 ⇒ 00:20:53.960 Amber Lin: since Awaish is still here. Anything you guys need help from Aish.
242 00:21:01.570 ⇒ 00:21:03.969 Amber Lin: Look at these cars.
243 00:21:03.970 ⇒ 00:21:12.860 Casie Aviles: I guess, for Dougster. Not, I guess, not a huge priority. But I was just wondering if
244 00:21:13.330 ⇒ 00:21:16.489 Casie Aviles: we we have access to that like.
245 00:21:16.720 ⇒ 00:21:19.550 Casie Aviles: do we do? We get invited to the platform.
246 00:21:21.011 ⇒ 00:21:27.150 Awaish Kumar: Yeah, I don’t know how to handle this. I think we should ask utham like it was a
247 00:21:28.990 ⇒ 00:21:36.239 Awaish Kumar: like the individual like the trial version. I I just use it using my
248 00:21:38.257 ⇒ 00:21:46.304 Awaish Kumar: own credentials, so I don’t know if I can invite. I I will see if there is any way to invite the team, I will invite you, and then
249 00:21:47.120 ⇒ 00:21:51.230 Awaish Kumar: otherwise, we have to figure out to have some common
250 00:21:51.640 ⇒ 00:21:55.620 Awaish Kumar: email where we can. All which we can all use to log in.
251 00:21:56.960 ⇒ 00:21:57.360 Amber Lin: Okay.
252 00:21:57.360 ⇒ 00:21:58.479 Awaish Kumar: So I will see.
253 00:22:02.550 ⇒ 00:22:03.240 Amber Lin: Yeah, okay.
254 00:22:03.240 ⇒ 00:22:03.710 Awaish Kumar: And.
255 00:22:03.710 ⇒ 00:22:04.820 Amber Lin: Input.
256 00:22:06.459 ⇒ 00:22:16.109 Amber Lin: Yeah, in in here, if it’s blocked or whatever just feel free to move it to either blocked or it needs help from.
257 00:22:16.110 ⇒ 00:22:17.290 Awaish Kumar: Okay. Yeah. Sure.
258 00:22:17.290 ⇒ 00:22:21.430 Amber Lin: Yeah, how is the these 2? So.
259 00:22:21.430 ⇒ 00:22:32.120 Awaish Kumar: Right for 1 8. I already inform you that it is okay. I was talking about other one here.
260 00:22:32.790 ⇒ 00:22:37.340 Awaish Kumar: So it it is like kind of zoom about zoom scripts which basically the
261 00:22:37.940 ⇒ 00:22:44.129 Awaish Kumar: I I mentioned that like right now, I haven’t been. I spent some time initially
262 00:22:44.340 ⇒ 00:22:48.860 Awaish Kumar: to look for ways, but after that I just worked on other sources and left this.
263 00:22:49.130 ⇒ 00:22:58.419 Awaish Kumar: And I didn’t touch this one, because we already have zoom data in the aws! And the the team is already progressing on that.
264 00:22:58.540 ⇒ 00:23:04.470 Awaish Kumar: So I just kept it as a blocked one. So we need to explore Dexter further.
265 00:23:04.790 ⇒ 00:23:09.649 Awaish Kumar: Like use all its features, to to see if it is possible with the text or not.
266 00:23:10.226 ⇒ 00:23:16.830 Awaish Kumar: With my current knowledge I haven’t been able to find out a straightforward way to do handle that
267 00:23:17.430 ⇒ 00:23:19.179 Awaish Kumar: to end like the these
268 00:23:19.390 ⇒ 00:23:28.319 Awaish Kumar: asynchronous events. So yeah, zoom, basically with the dexter we choose like schedule based. Etl runs.
269 00:23:28.440 ⇒ 00:23:33.240 Awaish Kumar: I’m not sure how it will handle this event streaming.
270 00:23:33.440 ⇒ 00:23:36.409 Awaish Kumar: So this was kind of blocked, not in progress.
271 00:23:38.470 ⇒ 00:23:42.779 Amber Lin: Blocks because we need to explore
272 00:23:50.070 ⇒ 00:23:52.410 Amber Lin: So let’s see.
273 00:23:53.740 ⇒ 00:23:57.930 Awaish Kumar: So like I can spend some more time on looking.
274 00:23:58.760 ⇒ 00:24:00.479 Awaish Kumar: So number. One thing is that.
275 00:24:00.480 ⇒ 00:24:00.930 Amber Lin: What?
276 00:24:00.930 ⇒ 00:24:01.420 Awaish Kumar: We have.
277 00:24:01.420 ⇒ 00:24:04.390 Amber Lin: I’m sorry I I got.
278 00:24:04.715 ⇒ 00:24:05.040 Awaish Kumar: So.
279 00:24:05.040 ⇒ 00:24:05.710 Amber Lin: Correct.
280 00:24:06.320 ⇒ 00:24:15.229 Awaish Kumar: So so number one thing is that we have a source zoom from where we want to move migrate data from Zoom Platform to
281 00:24:15.903 ⇒ 00:24:24.060 Awaish Kumar: to the S. 3, and we had windmill already built like set up by Casey and Meg to
282 00:24:24.430 ⇒ 00:24:32.939 Awaish Kumar: to to move that data. They already moved the data to S. 3. And they already built in and built an agent, I think, based on that
283 00:24:33.546 ⇒ 00:24:46.240 Awaish Kumar: or build build some embeddings and everything. But yeah, the thing is, we wanted to have a a platform which we can use to basically a single platform where we can use.
284 00:24:46.370 ⇒ 00:24:50.569 Awaish Kumar: which we can use to migrate all these sources. So right now
285 00:24:51.512 ⇒ 00:24:58.830 Awaish Kumar: with the Dagster, we, the we can handle like schedule base. Etl runs.
286 00:24:59.540 ⇒ 00:25:03.179 Awaish Kumar: and I’m not sure how it is going to handle event based
287 00:25:04.406 ⇒ 00:25:11.244 Awaish Kumar: sources. So we are using Dexter for Gita pipeline right now and
288 00:25:12.560 ⇒ 00:25:19.799 Awaish Kumar: and and because that is like a scheduled one. It runs once a day to move migrate data, but for the zoom.
289 00:25:19.920 ⇒ 00:25:21.540 Awaish Kumar: are we?
290 00:25:21.770 ⇒ 00:25:31.000 Awaish Kumar: I just put it as blocked because we already had something for the zoom which was working, so I can spend some more time if needed right now, or maybe in a
291 00:25:31.510 ⇒ 00:25:37.869 Awaish Kumar: in after some time, because I’m also new to Dagester. I have to learn all its features, and
292 00:25:37.990 ⇒ 00:25:42.430 Awaish Kumar: and it’s like all it’s on on what it can do.
293 00:25:42.600 ⇒ 00:25:48.829 Amber Lin: Dexter is, that is, that the platform that we all decided to use? Is this it.
294 00:25:50.794 ⇒ 00:25:57.909 Awaish Kumar: Yes, Dexter is an orchestration tool. So we have multiple sources, and we need some tool to handle that.
295 00:25:58.240 ⇒ 00:25:58.800 Amber Lin: Using.
296 00:25:58.800 ⇒ 00:25:59.420 Awaish Kumar: Extra.
297 00:25:59.420 ⇒ 00:26:06.160 Amber Lin: We’re not. We decided to check right? No other options. We’re just gonna use this.
298 00:26:07.100 ⇒ 00:26:08.549 Awaish Kumar: So, yeah, we had.
299 00:26:08.550 ⇒ 00:26:09.110 Amber Lin: Make it sounds.
300 00:26:09.110 ⇒ 00:26:19.410 Awaish Kumar: 2 options which we discussed with Utam like airflow and the the Prefect, and and but we decided to go with Dexter.
301 00:26:20.170 ⇒ 00:26:28.530 Amber Lin: I see? Megan Casey, do we need this? Urgently, or can this be sort of like running in the back.
302 00:26:29.020 ⇒ 00:26:31.109 Casie Aviles: No, it’s it can be running in the back.
303 00:26:31.110 ⇒ 00:26:31.839 Miguel de Veyra: Yeah, it can be done.
304 00:26:33.390 ⇒ 00:26:39.670 Casie Aviles: Yeah, because for the event based scripts that we have as a wish mentioned, we have that on windmill
305 00:26:39.830 ⇒ 00:26:46.250 Casie Aviles: and yeah, bottom. Re, re, I think he already set it up again and paid for it.
306 00:26:47.180 ⇒ 00:26:54.870 Amber Lin: So I’ll move it to ready for development. I’ll create a separate ticket just for Daxter discovery.
307 00:26:56.330 ⇒ 00:26:57.950 Amber Lin: Spike.
308 00:26:58.370 ⇒ 00:26:59.630 Amber Lin: It’s a company.
309 00:27:02.940 ⇒ 00:27:08.079 Amber Lin: Yeah. And I will put this as like a medium or low priority.
310 00:27:09.020 ⇒ 00:27:10.090 Amber Lin: Hmm!
311 00:27:14.010 ⇒ 00:27:19.729 Amber Lin: Do we want to explore Daxter? This cycle or next cycle?
312 00:27:23.360 ⇒ 00:27:26.559 Awaish Kumar: Like, it’s okay. I can do this in cycle in this cycle.
313 00:27:27.742 ⇒ 00:27:34.339 Amber Lin: I I know you have a lot of other things to do. So I just wanna know, like, for all question for all 3 of you. Do we want to do.
314 00:27:34.340 ⇒ 00:27:39.309 Awaish Kumar: Yeah, right now, I think I have the capability, like the availability to work on this. Yeah.
315 00:27:39.520 ⇒ 00:27:42.690 Amber Lin: Okay, so I’ll just put it as a discovery.
316 00:27:43.747 ⇒ 00:27:47.800 Amber Lin: How long do you think this will take like somewhere around here?
317 00:27:47.800 ⇒ 00:27:55.073 Awaish Kumar: Not sure. Like, I’m I’m just learning this new tool. So I’m okay. So so sure, yeah.
318 00:27:55.930 ⇒ 00:28:01.369 Amber Lin: Okay, what is the goal? We want to get out of that learning sort of like. If.
319 00:28:01.370 ⇒ 00:28:02.110 Awaish Kumar: Yeah.
320 00:28:02.110 ⇒ 00:28:02.770 Amber Lin: We could see.
321 00:28:02.770 ⇒ 00:28:14.070 Awaish Kumar: Like Dexter is a kind of a general tool. It’s not specific to AI. So I’m learning some of its features like number one like. If cater can handle event based.
322 00:28:14.470 ⇒ 00:28:25.969 Awaish Kumar: Pipeline number 2 is, if if it can, how it can handle like Dbt runs, how it can trigger like runs from 5 trend or polytomic.
323 00:28:26.230 ⇒ 00:28:28.680 Awaish Kumar: So this is kind of
324 00:28:29.590 ⇒ 00:28:34.765 Awaish Kumar: something which is for useful for all the teams. Not just not just the I.
325 00:28:35.320 ⇒ 00:28:43.580 Amber Lin: I see. I see, I see. So okay, so I will. Just
326 00:28:49.120 ⇒ 00:29:00.710 Amber Lin: So I guess for us. The only question we’ll ask you is that, can they handle events based pipelines? And I guess we’ll like for our next, or maybe, like midweek.
327 00:29:01.080 ⇒ 00:29:04.029 Amber Lin: we’ll check in with you. How that goes. How does that.
328 00:29:04.030 ⇒ 00:29:04.620 Awaish Kumar: Yep.
329 00:29:04.810 ⇒ 00:29:07.069 Amber Lin: Okay. So I’ll say, like Wednesday
330 00:29:07.410 ⇒ 00:29:09.140 Amber Lin: or Tuesday, what do you think.
331 00:29:12.050 ⇒ 00:29:12.690 Awaish Kumar: Like
332 00:29:13.740 ⇒ 00:29:15.340 Awaish Kumar: Wednesdays would be fine.
333 00:29:16.560 ⇒ 00:29:17.439 Amber Lin: Oh, pardon me.
334 00:29:18.610 ⇒ 00:29:20.800 Awaish Kumar: Sorry, I said. Wednesday is is okay.
335 00:29:20.800 ⇒ 00:29:22.300 Amber Lin: Okay. Wednesday
336 00:29:27.980 ⇒ 00:29:29.380 Amber Lin: sounds good.
337 00:29:32.640 ⇒ 00:29:34.920 Amber Lin: Casey amigo. Is that everything that.
338 00:29:34.920 ⇒ 00:29:35.500 Awaish Kumar: I’m being.
339 00:29:35.500 ⇒ 00:29:41.799 Amber Lin: From away from any other data stuff that we wanted help with.
340 00:29:41.800 ⇒ 00:29:42.939 Awaish Kumar: So like.
341 00:29:43.680 ⇒ 00:29:52.020 Awaish Kumar: I can also look at like, if you want to move data from s. 3 to super base, or if we want to
342 00:29:52.180 ⇒ 00:29:58.440 Awaish Kumar: do some what you say. Data transformation
343 00:29:58.620 ⇒ 00:30:07.209 Awaish Kumar: while moving data from S 3 to super base. So like, I’m up for all kind, all all of these things, because I have.
344 00:30:07.630 ⇒ 00:30:09.699 Awaish Kumar: and like
345 00:30:10.150 ⇒ 00:30:19.198 Awaish Kumar: I’ve got message from both of them, like I kind of got a feedback from both of them that like I can be if anywhere needed. I can be like
346 00:30:19.900 ⇒ 00:30:25.690 Awaish Kumar: be health if I can be helpful anywhere in the in the process, like I can take it.
347 00:30:27.190 ⇒ 00:30:27.830 Casie Aviles: Okay.
348 00:30:30.140 ⇒ 00:30:41.329 Awaish Kumar: So obviously, I can. I haven’t worked on the agents itself. So yeah, that’s your expertise. But I can help with data transformation or moving data from S. 3 to super base, or anywhere like that.
349 00:30:45.750 ⇒ 00:30:46.340 Amber Lin: Don’t you guys.
350 00:30:46.808 ⇒ 00:30:48.680 Miguel de Veyra: We’ll let you know.
351 00:30:49.110 ⇒ 00:30:54.850 Miguel de Veyra: Yeah, because right now, our priority really is to get the an agent out. But definitely.
352 00:30:54.850 ⇒ 00:30:57.149 Amber Lin: Yeah, I mean, if you.
353 00:30:57.460 ⇒ 00:31:13.680 Amber Lin: since we’re we are doing some formatting with the small batch, right? We are doing it for matter more. And we are experimenting with the different formats. Do you guys want to work with a ways on deciding like having a few
354 00:31:14.080 ⇒ 00:31:17.329 Amber Lin: few ways we can do it, or just help getting help
355 00:31:17.917 ⇒ 00:31:25.920 Amber Lin: or even just getting a wish up to speed of what we’re actually doing here. What do you guys think so. I can just have
356 00:31:26.090 ⇒ 00:31:28.130 Amber Lin: and just add it to the amount of more.
357 00:31:28.600 ⇒ 00:31:30.969 Amber Lin: If you guys want to work on it together.
358 00:31:31.530 ⇒ 00:31:33.749 Casie Aviles: Okay, yeah, I guess
359 00:31:35.080 ⇒ 00:31:42.089 Casie Aviles: so. We could ask for his like thoughts. Wish I could. We could ask for your thoughts on how we could move
360 00:31:43.196 ⇒ 00:31:47.189 Casie Aviles: the data from S, 3. As you can see, it’s in 4 K file. So
361 00:31:47.590 ⇒ 00:31:49.620 Casie Aviles: something we haven’t worked on yet.
362 00:31:51.310 ⇒ 00:31:54.640 Casie Aviles: And yeah, so we’re not sure how we could exactly
363 00:31:55.030 ⇒ 00:32:00.440 Casie Aviles: like, what’s the best way to structure it? Do we need to transform it like, can we filter it down to
364 00:32:02.050 ⇒ 00:32:06.019 Casie Aviles: to like a specific channel, a specific client, something like that.
365 00:32:07.780 ⇒ 00:32:08.485 Awaish Kumar: Okay?
366 00:32:09.520 ⇒ 00:32:17.039 Awaish Kumar: so yeah, like, it really depends. What is your requirement like how you want to feed it to the agent.
367 00:32:17.150 ⇒ 00:32:20.190 Awaish Kumar: So for example, if we want to have
368 00:32:21.525 ⇒ 00:32:25.330 Awaish Kumar: want to filter the data for the by the client, so we might
369 00:32:25.530 ⇒ 00:32:28.670 Awaish Kumar: see that there are some specific channels
370 00:32:28.920 ⇒ 00:32:34.264 Awaish Kumar: for this specific client. We only need data from these channels.
371 00:32:36.025 ⇒ 00:32:43.790 Awaish Kumar: for this client. And then maybe in general channels, we can look for keywords
372 00:32:43.890 ⇒ 00:32:50.690 Awaish Kumar: for this client and get those messages only in in our in our data set for the thank you.
373 00:32:51.060 ⇒ 00:32:55.129 Awaish Kumar: For this specific agent. So these are kind of 2 ways.
374 00:32:55.380 ⇒ 00:33:04.309 Awaish Kumar: But for this, like, obviously, we need some data transformation. Because as I already mentioned, the the data in the slack is
375 00:33:05.880 ⇒ 00:33:08.459 Awaish Kumar: data which is coming from slack is
376 00:33:08.590 ⇒ 00:33:28.269 Awaish Kumar: is not entirely how we have structured it. It’s kind of coming from the polytomic. So yeah, we what we can do is that we can decide on. Okay, how? Like what? Number one exploring the data. So you have. These parquet files are like, you have to re like, you cannot
377 00:33:28.930 ⇒ 00:33:33.790 Awaish Kumar: directly read it, but you can maybe read it through Ennet editn, or
378 00:33:34.724 ⇒ 00:33:45.105 Awaish Kumar: or like the pandas, like in the Jupyter load books. You can. You can visualize these these files because, like Csv files are kind of human readable. But they are like,
379 00:33:45.650 ⇒ 00:33:52.449 Awaish Kumar: not compressed. So they are not good for basically keeping the data and also moving moving data from.
380 00:33:52.890 ⇒ 00:34:08.630 Awaish Kumar: So I kept it as a parquet because they are better compressed version of keeping the big data files. But what we can do is that we can try it out in. And the Jupyter notebooks
381 00:34:09.202 ⇒ 00:34:34.709 Awaish Kumar: to to like discover the data, how it is structured. And then they agree on the format. Like, like, okay, we we have in this format. But we need this data into this format to feed into this agent, so we can write our script in between. Then, if it is possible to do some transformation on net, and we can do there. Otherwise we can just write some python scripts, read data from this parquet files and
382 00:34:34.900 ⇒ 00:34:38.779 Awaish Kumar: put it back in s. 3. But in a in the format we want.
383 00:34:38.889 ⇒ 00:34:40.200 Awaish Kumar: So this is kind of
384 00:34:41.790 ⇒ 00:34:42.340 Casie Aviles: Okay.
385 00:34:42.719 ⇒ 00:34:52.679 Awaish Kumar: Data transform. This is the data transformation stage. But that really depends on how you want to structure the data for your agent like.
386 00:34:53.280 ⇒ 00:34:55.320 Awaish Kumar: So like, I’m not
387 00:34:55.710 ⇒ 00:35:02.459 Awaish Kumar: sure. Like I, if I want to build agent for a specific client, how would I structure my files?
388 00:35:02.610 ⇒ 00:35:10.170 Awaish Kumar: If, if, like, yeah, we can brainstorm on that. And then we can go for data transformation step.
389 00:35:10.900 ⇒ 00:35:13.090 Amber Lin: Yay, okay, okay.
390 00:35:13.804 ⇒ 00:35:29.980 Amber Lin: I think the matamor Asian will be pretty. We’ll just take a 1st version and just chunk out a matamor Asian to experiment with, because there’s very, very, very little data in there. But I think this experimentation should go on.
391 00:35:30.320 ⇒ 00:35:39.320 Amber Lin: When do you think is a good date?
392 00:35:40.830 ⇒ 00:35:41.769 Awaish Kumar: Yeah. But
393 00:35:44.730 ⇒ 00:35:52.490 Awaish Kumar: yeah. But like, as I mentioned, I I need, I will need this input from the other members on how they want
394 00:35:52.870 ⇒ 00:36:00.119 Awaish Kumar: be structured for AI interm. If we think in terms of like an Llm. Agent like how.
395 00:36:00.120 ⇒ 00:36:00.720 Amber Lin: Hmm.
396 00:36:00.720 ⇒ 00:36:02.480 Awaish Kumar: How you want to feed it like.
397 00:36:02.630 ⇒ 00:36:10.020 Awaish Kumar: then I can discover like, how can we achieve this transformation? So there are 2 steps of doing this.
398 00:36:12.840 ⇒ 00:36:13.460 Casie Aviles: Okay.
399 00:36:14.570 ⇒ 00:36:19.290 Casie Aviles: So yeah, I, I, I still need to like figure out that
400 00:36:19.720 ⇒ 00:36:23.889 Casie Aviles: that part like with, like, how exactly we want to structure it.
401 00:36:24.640 ⇒ 00:36:30.699 Casie Aviles: So I I guess I could just, you know, ping a wish if I have additional questions.
402 00:36:31.900 ⇒ 00:36:32.410 Amber Lin: Should I make.
403 00:36:32.410 ⇒ 00:36:33.950 Awaish Kumar: Yeah, sure, no worries.
404 00:36:34.390 ⇒ 00:36:38.590 Awaish Kumar: You can just ping me and on slack.
405 00:36:38.980 ⇒ 00:36:45.180 Awaish Kumar: and I can hop on a call if if I I’m available, or I can just reply.
406 00:36:45.410 ⇒ 00:36:47.819 Awaish Kumar: but yeah, obviously, you can reach out to me.
407 00:36:49.080 ⇒ 00:36:53.179 Casie Aviles: Sure. Thank you. Yeah, because we already have like an idea how we
408 00:36:53.400 ⇒ 00:36:55.869 Casie Aviles: want to do it. So I think that’s what
409 00:36:56.140 ⇒ 00:36:58.880 Casie Aviles: Miguel suggested. So we’ll try those first.st
410 00:36:59.310 ⇒ 00:36:59.840 Miguel de Veyra: I think.
411 00:36:59.840 ⇒ 00:37:00.690 Casie Aviles: Yeah.
412 00:37:01.080 ⇒ 00:37:15.710 Miguel de Veyra: I think the one we cause. I think the one we just we think we’ll do. Casey is good, like I I’d stick with that for the new messages coming in for the new messages trickling down. But the problem now is the ones that existed before.
413 00:37:17.060 ⇒ 00:37:17.940 Casie Aviles: For the past.
414 00:37:18.230 ⇒ 00:37:18.730 Casie Aviles: That’s true.
415 00:37:18.730 ⇒ 00:37:21.950 Miguel de Veyra: That. Yeah, I think that’s where we’ll need this.
416 00:37:22.490 ⇒ 00:37:29.979 Casie Aviles: Yeah, that’s why the polytomic part is important, because we have everything there, basically. And we don’t have to
417 00:37:30.390 ⇒ 00:37:31.930 Casie Aviles: backfill anymore.
418 00:37:32.130 ⇒ 00:37:42.999 Miguel de Veyra: Yeah, I think what we could do is just basically do what what we plan to do now and show it always. Just so it’s easier to communicate right? So he can see
419 00:37:43.390 ⇒ 00:37:44.150 Miguel de Veyra: basically works.
420 00:37:45.420 ⇒ 00:37:59.739 Awaish Kumar: Yeah, my, my question is like you. If you are making a agent specific to a client called metamor, then this the data from slack. If it’s it’s like thread, then it will have all the threads for all the clients.
421 00:37:59.850 ⇒ 00:38:08.079 Awaish Kumar: So basically how you are going to filter it. If you can, you are able to do it dynamically, then it’s okay. Otherwise, if you want it to be, have a
422 00:38:08.380 ⇒ 00:38:15.140 Awaish Kumar: in a separate file, then we can add in that stage in our data transformation. And
423 00:38:15.730 ⇒ 00:38:33.130 Awaish Kumar: the way I would suggest, like, simplest way would be that if you do a data discovery and we can create a simple excel sheet kind of thing where you can say, Okay, this is how it looks. This is how we want it. And we are able to do it, or we need to help. So yeah.
424 00:38:34.000 ⇒ 00:38:45.139 Miguel de Veyra: Yeah. Cause. One of the things we have right now wish is that we have this thing called. Basically it tracks the messages for client channels. If you know people are communicating in it or not.
425 00:38:45.980 ⇒ 00:39:00.210 Miguel de Veyra: So we basically have access to this messages to the daily messages. Or I think we can also do it weekly, Casey, right? But yeah, for now we do have the weekly, either daily messages. So what we’re planning to do is basically to use those.
426 00:39:00.720 ⇒ 00:39:12.449 Miguel de Veyra: but to get those daily messages and just throw them and embed them into slack directly, without having to depend on S. 3. Because in that that way, we can just extend basically the workflow.
427 00:39:14.830 ⇒ 00:39:17.760 Awaish Kumar: Oh, yeah, but but like.
428 00:39:18.010 ⇒ 00:39:23.299 Awaish Kumar: okay. But like with the polytomic, you have data from all the channels and all the messages.
429 00:39:25.695 ⇒ 00:39:31.499 Miguel de Veyra: But I don’t think we really need all the channels. It’s specific channels, only that we need to prioritize.
430 00:39:31.500 ⇒ 00:39:54.639 Awaish Kumar: But yeah, that’s that’s we that that’s my the thing about data transformation which I talked is is about that, like, we have data for all the channels. But you only care about meta mode, for example, for now. So how can we get that data from the the S. 3 and feed it to to our agent. So yeah, that’s that’s where I want to work.
431 00:39:54.940 ⇒ 00:39:58.060 Awaish Kumar: So it is scalable. So like
432 00:39:58.290 ⇒ 00:40:04.200 Awaish Kumar: for all the clients. So once we discover that then that same flow will work for every client.
433 00:40:05.180 ⇒ 00:40:06.140 Miguel de Veyra: Yes, yes.
434 00:40:10.830 ⇒ 00:40:18.009 Amber Lin: Awesome! So how fast would this go?
435 00:40:18.870 ⇒ 00:40:21.879 Amber Lin: So when can we have? When can we have that done.
436 00:40:24.808 ⇒ 00:40:27.859 Miguel de Veyra: That’s a lot of tickets. I think that we need to break it down.
437 00:40:30.205 ⇒ 00:40:30.990 Amber Lin: where.
438 00:40:30.990 ⇒ 00:40:31.340 Miguel de Veyra: So.
439 00:40:31.340 ⇒ 00:40:33.170 Amber Lin: Need to break down tickets.
440 00:40:33.690 ⇒ 00:40:37.679 Miguel de Veyra: Yeah. So I think one of the wait, I’m thinking
441 00:40:40.280 ⇒ 00:40:54.269 Miguel de Veyra: 1 1 of the tickets would probably be definitely, I think we should actually just start with, it is basically loading slack data from the part that parquet files into basically how we can transform that.
442 00:40:55.740 ⇒ 00:40:56.260 Amber Lin: Can you.
443 00:40:56.260 ⇒ 00:40:56.950 Awaish Kumar: Yeah, so.
444 00:40:56.950 ⇒ 00:40:57.455 Amber Lin: Yeah.
445 00:40:59.920 ⇒ 00:41:01.170 Miguel de Veyra: Okay, sure. I can add that to them.
446 00:41:01.170 ⇒ 00:41:04.160 Amber Lin: Yeah, I am missing keywords.
447 00:41:06.660 ⇒ 00:41:12.290 Miguel de Veyra: How to load slacker block data from S. 3.
448 00:41:12.690 ⇒ 00:41:13.310 Amber Lin: Hmm.
449 00:41:16.670 ⇒ 00:41:19.539 Miguel de Veyra: How to transform, how to load and transform.
450 00:41:24.310 ⇒ 00:41:25.989 Miguel de Veyra: I’ll assign this to a wish.
451 00:41:26.540 ⇒ 00:41:27.210 Amber Lin: The.
452 00:41:28.220 ⇒ 00:41:34.909 Miguel de Veyra: And then I think, Judith, I’m not sure
453 00:41:36.050 ⇒ 00:41:40.909 Miguel de Veyra: I’ll just create it, for now added it in requirement started.
454 00:41:44.770 ⇒ 00:41:45.360 Miguel de Veyra: Yeah.
455 00:41:45.690 ⇒ 00:41:50.420 Awaish Kumar: So can can any of you help me with like creating a sample sheet for me
456 00:41:50.550 ⇒ 00:42:00.780 Awaish Kumar: that like that like, just read that. Read the data in Pake, like simple data and see like
457 00:42:01.780 ⇒ 00:42:11.490 Awaish Kumar: like the this is the format we have the data in. And this is the format you you wanted to be in a final stage. So just like maybe 2 sem 2 sheets with
458 00:42:11.660 ⇒ 00:42:15.579 Awaish Kumar: within like maybe 1015 rows, it’s enough.
459 00:42:16.460 ⇒ 00:42:17.170 Casie Aviles: Okay.
460 00:42:17.170 ⇒ 00:42:27.380 Amber Lin: Hmm, so we should create, create, spread sample spreadsheets.
461 00:42:27.740 ⇒ 00:42:34.079 Amber Lin: How we want to see data is that how it is.
462 00:42:35.520 ⇒ 00:42:37.319 Awaish Kumar: Yes, yes. Kind of.
463 00:42:40.180 ⇒ 00:42:42.009 Amber Lin: How? What would you need?
464 00:42:42.490 ⇒ 00:42:43.760 Amber Lin: What do you need?
465 00:42:45.130 ⇒ 00:42:47.650 Casie Aviles: From us. He would need the
466 00:42:48.760 ⇒ 00:42:56.129 Casie Aviles: the transfer, the trend like the goal, like the the how the data would need to be to look like, and the transform.
467 00:43:03.140 ⇒ 00:43:05.680 Miguel de Veyra: You mean like the columns, and shit right.
468 00:43:06.960 ⇒ 00:43:07.730 Casie Aviles: Yeah, like.
469 00:43:08.270 ⇒ 00:43:09.730 Miguel de Veyra: Okay, okay, yeah, that makes sense.
470 00:43:09.730 ⇒ 00:43:13.094 Awaish Kumar: And like, like what you need like. For example, you need
471 00:43:13.660 ⇒ 00:43:24.780 Awaish Kumar: messages, and you also need threads. But, like, in what format do you need? That’s the question. So you just create simple, like, we have the data already. You just copy paste from there.
472 00:43:24.960 ⇒ 00:43:31.360 Awaish Kumar: create some samples here and there, and then I can figure out, okay, how to achieve that.
473 00:43:32.650 ⇒ 00:43:33.580 Miguel de Veyra: Okay. Okay.
474 00:43:33.580 ⇒ 00:43:34.190 Amber Lin: Hmm!
475 00:43:35.160 ⇒ 00:43:38.349 Amber Lin: So when you say what format it would be like.
476 00:43:38.950 ⇒ 00:43:43.500 Awaish Kumar: Yeah, format is the schema like structure of the file.
477 00:43:43.750 ⇒ 00:43:46.859 Awaish Kumar: Yeah, columns needed things like that.
478 00:43:49.430 ⇒ 00:43:52.639 Amber Lin: Okay, I think this is something we can do today. Right?
479 00:43:58.902 ⇒ 00:44:05.390 Miguel de Veyra: Probably not today, because we want to work on the agent. And then there’s like ABC stuff that needs to be worked on.
480 00:44:06.110 ⇒ 00:44:08.610 Amber Lin: Is there ABC, stuff that needs to be worked on.
481 00:44:10.620 ⇒ 00:44:13.420 Miguel de Veyra: I think. Wait! Let me double check.
482 00:44:13.420 ⇒ 00:44:16.080 Amber Lin: I don’t think we’ve assigned anything, and it’s.
483 00:44:16.080 ⇒ 00:44:18.309 Miguel de Veyra: Oh, yeah, yeah, we put it in in.
484 00:44:18.310 ⇒ 00:44:18.700 Amber Lin: Yeah.
485 00:44:18.700 ⇒ 00:44:20.070 Miguel de Veyra: Requirements started.
486 00:44:20.670 ⇒ 00:44:24.349 Amber Lin: Yeah. Hey? I think.
487 00:44:24.350 ⇒ 00:44:26.520 Miguel de Veyra: I think I can. I can work on this today.
488 00:44:26.520 ⇒ 00:44:29.609 Miguel de Veyra: Yeah, let’s so we need to discuss this anyways with Casey.
489 00:44:29.610 ⇒ 00:44:34.170 Amber Lin: Yeah, let’s have this so that a wish can get started because.
490 00:44:34.170 ⇒ 00:44:35.490 Miguel de Veyra: This is probably a 3.
491 00:44:35.490 ⇒ 00:44:37.190 Amber Lin: Thing right?
492 00:44:41.350 ⇒ 00:44:45.950 Amber Lin: Right, that’s good. So I would say that.
493 00:44:52.750 ⇒ 00:44:56.230 Amber Lin: wish. Do you have to wait until
494 00:44:56.480 ⇒ 00:45:02.729 Amber Lin: the document is there, or is there like other things you can be looking at.
495 00:45:03.210 ⇒ 00:45:07.624 Awaish Kumar: Yeah, I can. I have also to look at how the slack data and
496 00:45:09.331 ⇒ 00:45:14.470 Awaish Kumar: I also have to see all like that. I have to do the data discovery also.
497 00:45:15.060 ⇒ 00:45:18.560 Awaish Kumar: Finally, like, when I have the document, then I can work on it.
498 00:45:19.920 ⇒ 00:45:28.430 Amber Lin: I see. So you can. You say you can start with the discovery now and then? Wait. Once the document is there, you can start the transformation.
499 00:45:28.940 ⇒ 00:45:29.770 Awaish Kumar: Yes.
500 00:45:32.470 ⇒ 00:45:35.910 Amber Lin: Spain, old spreadsheet!
501 00:45:37.240 ⇒ 00:45:39.300 Amber Lin: Do you have access to all the data.
502 00:45:41.410 ⇒ 00:45:41.990 Awaish Kumar: Yes!
503 00:45:42.460 ⇒ 00:45:44.579 Amber Lin: Okay, sounds good.
504 00:45:49.470 ⇒ 00:45:55.600 Amber Lin: I might just create a where is it?
505 00:45:56.320 ⇒ 00:45:57.110 Amber Lin: Fun?
506 00:45:58.980 ⇒ 00:46:01.159 Amber Lin: I would say. This is blocked.
507 00:46:01.900 ⇒ 00:46:03.519 Amber Lin: We’re ready for development.
508 00:46:04.580 ⇒ 00:46:05.729 Awaish Kumar: It’s too exciting.
509 00:46:05.960 ⇒ 00:46:10.900 Amber Lin: 89 covering.
510 00:46:14.130 ⇒ 00:46:15.140 Amber Lin: Oh.
511 00:46:19.070 ⇒ 00:46:21.289 Amber Lin: like around here, would you say?
512 00:46:26.510 ⇒ 00:46:27.520 Awaish Kumar: I don’t know. How.
513 00:46:27.890 ⇒ 00:46:30.619 Awaish Kumar: How do you read these points like.
514 00:46:32.193 ⇒ 00:46:39.450 Amber Lin: I think one points like an hour, 2 points like 3 to 4 h, 3 points. It’s like.
515 00:46:39.910 ⇒ 00:46:40.690 Amber Lin: yeah, let me go.
516 00:46:40.690 ⇒ 00:46:42.079 Casie Aviles: 4 to 5 h.
517 00:46:42.470 ⇒ 00:46:43.480 Amber Lin: Yeah.
518 00:46:45.240 ⇒ 00:46:45.970 Amber Lin: Oh.
519 00:46:45.970 ⇒ 00:46:46.640 Awaish Kumar: Sorry.
520 00:46:47.980 ⇒ 00:46:48.790 Amber Lin: Yeah.
521 00:46:48.920 ⇒ 00:46:58.559 Amber Lin: 2 points is 2 to 3 h, 3 points is 4 to 5 h. 5 points is 6 to 8 and 8 points is a day and a half to 2 days.
522 00:46:59.480 ⇒ 00:47:00.839 Awaish Kumar: We can put 3 here.
523 00:47:00.840 ⇒ 00:47:08.229 Amber Lin: Okay, so this is something like we do today. It is this too much.
524 00:47:11.440 ⇒ 00:47:13.030 Awaish Kumar: Yeah, I can work on this today.
525 00:47:13.680 ⇒ 00:47:14.430 Amber Lin: Okay.
526 00:47:15.030 ⇒ 00:47:16.520 Awaish Kumar: Sounds good.
527 00:47:16.720 ⇒ 00:47:17.750 Awaish Kumar: Yes, it.
528 00:47:18.560 ⇒ 00:47:25.879 Amber Lin: Yeah, I’ll trust you with your knowledge of your own capacity. But just let let us know if if it’s too much stuff.
529 00:47:29.710 ⇒ 00:47:34.199 Amber Lin: Yeah, Nigel, I would say, the follow up agent. We can.
530 00:47:34.800 ⇒ 00:47:41.140 Amber Lin: If there’s a lot today, I’ll just set the duty for tomorrow. If there’s too much stuff today. How’s that.
531 00:47:41.740 ⇒ 00:47:46.409 Miguel de Veyra: Yeah, because I think we, I think we need to also
532 00:47:47.710 ⇒ 00:47:54.990 Miguel de Veyra: create some sort of node that basically is a slack format. I think we do? Do we have that Casey? Already?
533 00:47:55.260 ⇒ 00:47:57.249 Miguel de Veyra: A slack formatter node or not? Yet.
534 00:47:57.640 ⇒ 00:48:03.890 Casie Aviles: I mean for that specific workflow. No, but we we do have like a slack format or node in other workflows.
535 00:48:03.890 ⇒ 00:48:09.679 Miguel de Veyra: Yeah, I think we can just do that and copy paste because Utah messaged in AI team.
536 00:48:10.890 ⇒ 00:48:14.320 Miguel de Veyra: And then he wants something basically formatted, too.
537 00:48:17.100 ⇒ 00:48:21.810 Miguel de Veyra: But yeah, but yeah, I can. I can work on the sales agent, because
538 00:48:21.920 ⇒ 00:48:27.990 Miguel de Veyra: if if we don’t have that by tomorrow, it’s gonna be questions.
539 00:48:28.770 ⇒ 00:48:33.490 Amber Lin: Okay, sounds good. So
540 00:48:33.880 ⇒ 00:48:47.140 Amber Lin: today we have sales agent model, or the spreadsheet to give to wish and waste. Looking at a bit of Daxter and bit of stock data discovery. And then, I think for these, they’re like.
541 00:48:47.670 ⇒ 00:48:52.079 Amber Lin: we need these to complete any of these. So I think
542 00:48:53.340 ⇒ 00:48:56.969 Amber Lin: by tomorrow we’ll have the slack, and then we’ll have next steps
543 00:48:58.010 ⇒ 00:49:01.620 Amber Lin: for will be on progress for Madamore.
544 00:49:02.000 ⇒ 00:49:02.889 Amber Lin: How’s that?
545 00:49:04.880 ⇒ 00:49:05.590 Casie Aviles: Okay.
546 00:49:05.980 ⇒ 00:49:06.773 Amber Lin: Yeah, okay.
547 00:49:07.690 ⇒ 00:49:08.580 Awaish Kumar: Are you?
548 00:49:10.330 ⇒ 00:49:16.379 Awaish Kumar: So like, are you guys, do you guys will be working on building the agent for slack in the meantime? Or
549 00:49:16.530 ⇒ 00:49:19.519 Awaish Kumar: are you waiting on our select data discovery.
550 00:49:21.450 ⇒ 00:49:23.079 Miguel de Veyra: No, we’re gonna be building it.
551 00:49:23.080 ⇒ 00:49:37.569 Amber Lin: Yeah, for my understanding what we wish, what we why we want to do this one is because this is a very, very recent client. We have very little data in slack. So we can. We wanted to experiment with different ways to
552 00:49:38.011 ⇒ 00:49:54.569 Amber Lin: to do that. So that’s where you’ll be really helpful in giving like different ideas of, maybe we should do this format. Maybe we should do that format because we sometimes we don’t even know what formats is available, at least I don’t know. So this is the experimentation project.
553 00:49:56.540 ⇒ 00:49:57.750 Awaish Kumar: Okay. Yeah.
554 00:49:57.750 ⇒ 00:49:58.470 Amber Lin: Yeah.
555 00:49:59.720 ⇒ 00:50:00.290 Awaish Kumar: Thank you.
556 00:50:00.590 ⇒ 00:50:03.170 Amber Lin: Yeah. Anything else?
557 00:50:07.570 ⇒ 00:50:09.449 Amber Lin: Okay, yeah, I will.
558 00:50:09.450 ⇒ 00:50:09.890 Casie Aviles: That’s it.
559 00:50:09.890 ⇒ 00:50:15.170 Amber Lin: Working on these. I will ask you guys what the
560 00:50:15.390 ⇒ 00:50:19.910 Amber Lin: Brainford want is, if I can’t find it, I’ll add it to matter more.
561 00:50:21.070 ⇒ 00:50:21.760 Amber Lin: Both.
562 00:50:21.760 ⇒ 00:50:24.190 Miguel de Veyra: Yeah. And then one of the other things.
563 00:50:24.655 ⇒ 00:50:28.990 Miguel de Veyra: Sorry, Amber. One of the things is, do we have the general information anywhere for matter more.
564 00:50:32.070 ⇒ 00:50:32.830 Amber Lin: Hey?
565 00:50:33.695 ⇒ 00:50:34.380 Amber Lin: Oops!
566 00:50:34.570 ⇒ 00:50:45.039 Amber Lin: Add brave storage on to chat channels and give channel.
567 00:50:48.500 ⇒ 00:50:50.390 Amber Lin: How’s that? I will do that?
568 00:50:50.850 ⇒ 00:50:51.670 Miguel de Veyra: Okay. Yeah.
569 00:50:51.670 ⇒ 00:50:52.280 Amber Lin: Yeah.
570 00:50:54.330 ⇒ 00:50:55.080 Amber Lin: Great.
571 00:50:56.710 ⇒ 00:50:57.640 Amber Lin: Okay.
572 00:50:57.940 ⇒ 00:50:59.180 Amber Lin: Sounds good.
573 00:51:01.750 ⇒ 00:51:03.250 Miguel de Veyra: Okay. Thanks. Everyone.
574 00:51:03.730 ⇒ 00:51:07.380 Amber Lin: Thank you, everyone. This is a long meeting. I appreciate you all.
575 00:51:08.310 ⇒ 00:51:09.070 Casie Aviles: Thank you.
576 00:51:10.040 ⇒ 00:51:10.960 Amber Lin: Bye-bye.