Meeting Title: AI-Data Platform Team Standup Date: 2025-08-20 Meeting participants: Mustafa Raja, Casie Aviles, Samuel Roberts, Uttam Kumaran
WEBVTT
1 00:00:11.360 ⇒ 00:00:12.830 Samuel Roberts: Ayy.
2 00:00:12.830 ⇒ 00:00:13.500 Mustafa Raja: Hey.
3 00:00:14.850 ⇒ 00:00:15.920 Samuel Roberts: How’s it going?
4 00:00:16.510 ⇒ 00:00:18.040 Mustafa Raja: Good. How about you?
5 00:00:18.040 ⇒ 00:00:22.410 Samuel Roberts: Doing alright, doing alright. I was playing around with Copilot kit, trying to get it.
6 00:00:22.410 ⇒ 00:00:22.910 Mustafa Raja: Oh, yeah.
7 00:00:22.910 ⇒ 00:00:29.140 Samuel Roberts: tied to N8N. I think I made some progress, but it’s… It’s getting there.
8 00:00:31.240 ⇒ 00:00:32.409 Samuel Roberts: Anytime.
9 00:00:37.660 ⇒ 00:00:38.870 Samuel Roberts: Right.
10 00:00:40.740 ⇒ 00:00:45.320 Samuel Roberts: Let me… Yeah.
11 00:00:48.000 ⇒ 00:00:51.980 Samuel Roberts: Actually, do any of you guys have linear up? You want to get that going? I kind of have a bunch of stuff.
12 00:00:53.440 ⇒ 00:00:55.220 Samuel Roberts: ProPilot kits all over my stuff.
13 00:00:55.610 ⇒ 00:01:00.510 Uttam Kumaran: Yeah, it should be on my machine in, like, 10 minutes, but maybe Casey or Mustafa, if you want to run?
14 00:01:00.510 ⇒ 00:01:01.669 Samuel Roberts: Yeah, yeah, if you wanna….
15 00:01:01.670 ⇒ 00:01:08.599 Uttam Kumaran: I didn’t get a chance to watch the co-pilot kit, I just was up pretty late, so I just got up, like, recently, I need to go check it all out.
16 00:01:08.870 ⇒ 00:01:17.170 Samuel Roberts: No, you’re good, you’re good. It was… it’s a… it’s a super simple thing, it’s just, I was just showing that it… it ties to N8N, but, Casey shared the ABC…
17 00:01:17.560 ⇒ 00:01:19.420 Samuel Roberts: workflow, so I can hopefully
18 00:01:19.550 ⇒ 00:01:25.520 Samuel Roberts: get it a little more tied together. I’m assuming that’s what you were thinking when you, like, wanted it, demonstrating that.
19 00:01:25.520 ⇒ 00:01:29.840 Uttam Kumaran: Exactly, yeah, like, yeah, like, probably the biggest two…
20 00:01:29.960 ⇒ 00:01:34.349 Uttam Kumaran: things I feel like that would be unlocks is, one, that, and then two, like.
21 00:01:34.750 ⇒ 00:01:42.100 Uttam Kumaran: probably just work on the monorepo eventually, like, I feel like those two… but this one, I want to put this in front of a couple clients, and then…
22 00:01:42.200 ⇒ 00:01:50.810 Uttam Kumaran: if we have, like, boilerplate code on how to do it, I think it’s pretty straightforward on how to migrate our existing chats to this, too, right? So….
23 00:01:51.350 ⇒ 00:01:52.070 Samuel Roberts: Yeah.
24 00:01:52.070 ⇒ 00:01:52.710 Uttam Kumaran: once.
25 00:01:52.710 ⇒ 00:01:57.320 Samuel Roberts: Yeah, it’s a little weird, because it has, like, backend,
26 00:01:57.860 ⇒ 00:01:59.950 Samuel Roberts: what do they call it? Service?
27 00:02:00.060 ⇒ 00:02:06.770 Samuel Roberts: Providers, and so, like, it has a bunch of… pre-built adapters for, like.
28 00:02:06.880 ⇒ 00:02:11.600 Samuel Roberts: OpenAI and other things, so I kind of had Cursor just…
29 00:02:11.730 ⇒ 00:02:23.799 Samuel Roberts: help me build one that will talk to the N8N webhooks, and handle everything on the front end the same, and it seems to work okay. We might have to do a little bit of tweaking, depending on each workflow, but, …
30 00:02:24.000 ⇒ 00:02:28.820 Samuel Roberts: once I… now that I have that working, I can probably standardize it a little bit more, so….
31 00:02:28.820 ⇒ 00:02:29.450 Uttam Kumaran: Okay.
32 00:02:29.840 ⇒ 00:02:47.599 Uttam Kumaran: Yeah, for me, you know, my whole thing was, like, look, even if it’s, like, complicated, it’s… we’re on NNN, but now, if we’re like, hey, if for anything that touches Copilot kit, maybe it has to all be in code, like, it is a compromise we can make and we can consider. Okay.
33 00:02:47.980 ⇒ 00:02:49.000 Uttam Kumaran: not, like…
34 00:02:49.110 ⇒ 00:03:03.979 Uttam Kumaran: not married to N8N in terms of, like… but I do think it’s great for prototyping, and then moving it to a code-based environment probably works. There’s also… yeah, I mean, that’s… that’s probably where I’m at right now, but I would love to… once I see the results, I can kind of make a decision.
35 00:03:04.360 ⇒ 00:03:11.170 Samuel Roberts: Yeah, yeah, definitely. I can probably throw a couple things together then, so maybe I can see how it would work with other tools as well.
36 00:03:11.310 ⇒ 00:03:16.140 Samuel Roberts: Better, like, more… integrated with Copilot Kit and see if it’s worth it.
37 00:03:16.260 ⇒ 00:03:26.010 Samuel Roberts: But yeah, I can definitely… I’m gonna… now that I got it, like, tying, or talking to NA then, I can at least… because that was a little bit of a hurdle to get it to do the right…
38 00:03:26.270 ⇒ 00:03:36.390 Samuel Roberts: webhooks and stuff, because it likes to talk to its own stuff better, but now that I’ve got that, I should be able to make a little progress with our current workflows, but I think it is a little bit of an uphill line.
39 00:03:38.580 ⇒ 00:03:41.180 Samuel Roberts: But anyway… yeah.
40 00:03:44.090 ⇒ 00:03:47.980 Samuel Roberts: Alright, what are we loose right here, Casey? Cool, okay.
41 00:03:48.150 ⇒ 00:03:49.550 Casie Aviles: You guys see my screen?
42 00:03:50.150 ⇒ 00:03:50.590 Samuel Roberts: Yep.
43 00:03:50.590 ⇒ 00:03:51.140 Uttam Kumaran: Yeah.
44 00:03:51.610 ⇒ 00:04:01.550 Casie Aviles: Okay, so I’ll just start from the, top, and don’t, move for, move to the bottom. So, yeah, I… let’s start with ABC, …
45 00:04:02.050 ⇒ 00:04:07.830 Casie Aviles: For me, I just have this ticket that’s, … for review, so…
46 00:04:08.190 ⇒ 00:04:12.629 Casie Aviles: I have a bunch of spike tickets for you, I just tagged you there are some.
47 00:04:13.390 ⇒ 00:04:13.930 Samuel Roberts: Okay.
48 00:04:14.670 ⇒ 00:04:19.789 Casie Aviles: So, I just love some… Feedback on some of the stuff there.
49 00:04:20.550 ⇒ 00:04:21.719 Casie Aviles: But yeah, …
50 00:04:26.240 ⇒ 00:04:26.990 Casie Aviles: Okay.
51 00:04:27.560 ⇒ 00:04:29.040 Samuel Roberts: Yeah, no, you’re gonna keep going.
52 00:04:29.350 ⇒ 00:04:31.339 Casie Aviles: Okay. So, okay.
53 00:04:31.910 ⇒ 00:04:37.099 Casie Aviles: all the stuff, … Is there anything on ABC’s side that…
54 00:04:37.750 ⇒ 00:04:40.119 Casie Aviles: You’d like to give an update on, or…?
55 00:04:40.120 ⇒ 00:04:51.189 Mustafa Raja: Yeah, for the one that is in the, review was, to update the hard-coded, JSON that we already have in the, what’s it called?
56 00:04:51.460 ⇒ 00:04:59.290 Mustafa Raja: In Anytin, and the one that is in progress, I didn’t really get, get the time to work on it.
57 00:05:01.780 ⇒ 00:05:03.240 Casie Aviles: Oh, okay, okay, this one.
58 00:05:08.870 ⇒ 00:05:09.430 Mustafa Raja: Yep.
59 00:05:09.720 ⇒ 00:05:10.390 Casie Aviles: Alright.
60 00:05:11.230 ⇒ 00:05:15.419 Casie Aviles: Yeah, I guess, other than that, for me, there are just…
61 00:05:15.830 ⇒ 00:05:20.069 Casie Aviles: some other triage tickets coming in that I will take a look at.
62 00:05:20.270 ⇒ 00:05:26.149 Casie Aviles: Okay. … there are some, just, some minor stuff fixes to me.
63 00:05:26.150 ⇒ 00:05:32.909 Uttam Kumaran: Like, how long is it taking, you know, daily or weekly on, like, handling triage? Like, how’s it working so far?
64 00:05:34.830 ⇒ 00:05:40.669 Casie Aviles: I typically give it around Let’s say 2 hours, I believe.
65 00:05:41.220 ⇒ 00:05:48.530 Casie Aviles: Or, depending on the issue, like, if it’s just a quick… Oh, per week, ….
66 00:05:48.670 ⇒ 00:05:49.840 Uttam Kumaran: Per day, 2 hours.
67 00:05:50.700 ⇒ 00:05:56.610 Casie Aviles: Yeah, per day. Like, if I have to go and fix it, fix this, triage…
68 00:05:56.790 ⇒ 00:05:59.419 Casie Aviles: In a day, then I would give around 2 hours.
69 00:06:00.310 ⇒ 00:06:11.129 Uttam Kumaran: And then, how many, like, are all of those issues things that we have to fix on our side, or are they things that the client can fix, or maybe
70 00:06:11.450 ⇒ 00:06:14.570 Uttam Kumaran: Just give us the synopsis of, like, what the tickets are so far.
71 00:06:14.750 ⇒ 00:06:17.049 Casie Aviles: Yeah, it’s a mix of both.
72 00:06:17.430 ⇒ 00:06:24.740 Casie Aviles: So some of these, Amber would assign to Janiece, because it’s, like, a gap on their side, like…
73 00:06:25.440 ⇒ 00:06:29.969 Casie Aviles: And then, for some of these, I have to take a look and…
74 00:06:30.360 ⇒ 00:06:37.549 Casie Aviles: check if it’s a problem with our, … with our side, and that would be… for example, if we have to update, like.
75 00:06:38.040 ⇒ 00:06:44.160 Casie Aviles: The inspector sheet. There’s, like, a missing inspector or an incorrect inspector assignment, so…
76 00:06:44.410 ⇒ 00:06:46.410 Casie Aviles: I would have to go and update that.
77 00:06:47.570 ⇒ 00:06:51.630 Casie Aviles: And then if it’s, like, a prompt thing, like a behavior thing, where…
78 00:06:52.980 ⇒ 00:06:56.449 Casie Aviles: For example, Andy still asks for something, even if it’s…
79 00:06:56.590 ⇒ 00:07:00.980 Casie Aviles: Clear, given the response of the user, then…
80 00:07:01.160 ⇒ 00:07:03.060 Casie Aviles: That’s something I have to fix.
81 00:07:04.190 ⇒ 00:07:10.729 Uttam Kumaran: Okay. I’ll, … I’ll join grooming today, but yeah, I’m gonna get you out of that.
82 00:07:10.730 ⇒ 00:07:11.770 Samuel Roberts: Okay.
83 00:07:11.770 ⇒ 00:07:17.400 Uttam Kumaran: I’m gonna create… we’re gonna work today on, like, SLAs by… pipe, and the.
84 00:07:17.400 ⇒ 00:07:18.030 Casie Aviles: Hmm.
85 00:07:18.250 ⇒ 00:07:20.800 Uttam Kumaran: It’s either gotta be Amber, or…
86 00:07:20.930 ⇒ 00:07:35.789 Uttam Kumaran: we’re gonna think about some process where you’re not checking this every day. Like, for urgent issues, sure, but I think we basically agreed with the client on, like, 48 hours for, like.
87 00:07:36.210 ⇒ 00:07:37.999 Uttam Kumaran: These types of changes.
88 00:07:38.130 ⇒ 00:07:45.269 Uttam Kumaran: And then for really systematic changes, we need to spend… we need to… Lengthen the, the time.
89 00:07:45.400 ⇒ 00:07:51.379 Uttam Kumaran: So… I… I don’t… I think 2 hours a day is gonna be quite a lot for you to handle.
90 00:07:52.260 ⇒ 00:07:57.000 Uttam Kumaran: … So I’m gr… I’m glad we caught that. I think let’s…
91 00:07:57.250 ⇒ 00:08:02.549 Uttam Kumaran: let’s take that as a note today in grooming. I mean, we have a couple options. One is, like.
92 00:08:02.890 ⇒ 00:08:06.839 Uttam Kumaran: Amber and the client… ideally, the client handles
93 00:08:07.430 ⇒ 00:08:12.169 Uttam Kumaran: Triaging, and then for anything that comes onto our plate.
94 00:08:12.320 ⇒ 00:08:15.050 Uttam Kumaran: we have an… we have an SLA, so…
95 00:08:15.190 ⇒ 00:08:17.999 Uttam Kumaran: At least there’s, like, 2 days to complete it.
96 00:08:18.140 ⇒ 00:08:21.599 Uttam Kumaran: And then for stuff that’s, like, non-urgent.
97 00:08:21.910 ⇒ 00:08:28.220 Uttam Kumaran: it should be at least 3 days. So, I… I would say don’t worry about checking stuff today.
98 00:08:28.540 ⇒ 00:08:28.860 Casie Aviles: Okay.
99 00:08:28.860 ⇒ 00:08:40.020 Uttam Kumaran: wait until we talk in grooming, I want to buy you some time back, because handling this every day, you just become, like, a service desk, so I want to get… make sure that you can, …
100 00:08:40.370 ⇒ 00:08:42.730 Uttam Kumaran: You can batch these, like, maybe it’s…
101 00:08:42.940 ⇒ 00:08:47.039 Uttam Kumaran: check in on Monday, Wednesday, Friday, or it’s like.
102 00:08:47.620 ⇒ 00:08:54.779 Uttam Kumaran: Just wait until the tickets manifest and handle it when you can handle it, but having to spend 2 hours a day going through this is gonna be…
103 00:08:55.800 ⇒ 00:08:57.090 Uttam Kumaran: Really, really tough.
104 00:08:57.960 ⇒ 00:09:05.559 Casie Aviles: Yeah, okay. Yeah, I agree. Yeah, I just… for each day, like, I just go and check every time a new ticket comes in.
105 00:09:05.680 ⇒ 00:09:08.070 Casie Aviles: But yeah, that’ll be helpful.
106 00:09:08.370 ⇒ 00:09:11.309 Uttam Kumaran: Your time is too valuable for us, so I don’t want….
107 00:09:11.310 ⇒ 00:09:12.099 Samuel Roberts: It’s like….
108 00:09:12.100 ⇒ 00:09:18.699 Uttam Kumaran: sit there and do that. Also, like, we’re… we have a bunch of other clients we’re supporting, so I want to make sure that they get
109 00:09:18.840 ⇒ 00:09:25.319 Uttam Kumaran: somewhat equal time, if possible, so… Okay, so let’s keep that as a note for grooming. I will, …
110 00:09:25.920 ⇒ 00:09:30.199 Uttam Kumaran: I’ll send a note right now to ABC Channel so that Amber can add that to the agenda.
111 00:09:30.980 ⇒ 00:09:31.630 Casie Aviles: Okay.
112 00:09:34.650 ⇒ 00:09:36.240 Casie Aviles: Okay, cool.
113 00:09:37.450 ⇒ 00:09:40.450 Casie Aviles: Yeah, I’m going to move to AI internal.
114 00:09:41.280 ⇒ 00:09:45.319 Uttam Kumaran: And then the one thing on ABC is, there’s a spike on…
115 00:09:45.750 ⇒ 00:09:49.679 Uttam Kumaran: How do we create a database around the zip codes, right?
116 00:09:50.020 ⇒ 00:09:51.050 Casie Aviles: Yes, yes.
117 00:09:51.700 ⇒ 00:09:54.530 Uttam Kumaran: Okay, I’ll take a look at that, so….
118 00:09:54.950 ⇒ 00:09:58.270 Casie Aviles: Yeah, I have the Notion doc for that, so….
119 00:09:58.270 ⇒ 00:09:58.840 Uttam Kumaran: Great.
120 00:09:59.570 ⇒ 00:10:09.849 Uttam Kumaran: How has it been creating, like, these spikes and technical design docs? Like, are you using… are you guys using, like, AI? Is it a mix? Like, what’s the process been so far?
121 00:10:11.180 ⇒ 00:10:16.250 Casie Aviles: For me, I… I… what I do is, I typically just…
122 00:10:16.450 ⇒ 00:10:21.099 Casie Aviles: list down my thoughts, like, raw thoughts, so I don’t…
123 00:10:21.750 ⇒ 00:10:31.680 Casie Aviles: immediately have AI just generate it, and then I provide it to AI, and, like, I give it the template, but sometimes, like, I feel like there’s too much…
124 00:10:32.180 ⇒ 00:10:37.869 Casie Aviles: That it generates, so I just try to ask it to… Cut it down.
125 00:10:38.290 ⇒ 00:10:44.930 Casie Aviles: And, keep it lean, so that not all of it is, like, coming out of… Thin error, like…
126 00:10:45.410 ⇒ 00:10:51.340 Casie Aviles: you know, just… I wanted to be grounded by what I actually observed, or what I actually looked at, so…
127 00:10:52.110 ⇒ 00:10:55.190 Casie Aviles: That’s how I typically handle the spikes.
128 00:10:55.750 ⇒ 00:11:02.299 Casie Aviles: I tried… although they’re… what I’ve created so far are just still very… Short, I guess, but…
129 00:11:02.940 ⇒ 00:11:08.399 Casie Aviles: Yeah, I don’t want it to be long, but has a lot of stuff that I don’t really…
130 00:11:08.880 ⇒ 00:11:09.750 Casie Aviles: You know.
131 00:11:09.890 ⇒ 00:11:10.790 Casie Aviles: Yeah.
132 00:11:11.610 ⇒ 00:11:12.580 Uttam Kumaran: Okay, okay.
133 00:11:13.250 ⇒ 00:11:18.149 Uttam Kumaran: How about everyone else? Like, how’s everyone doing with, like, creating spikes, or…
134 00:11:18.460 ⇒ 00:11:20.640 Uttam Kumaran: Using AI to augment that process.
135 00:11:20.640 ⇒ 00:11:26.670 Mustafa Raja: Yeah, my process would be, quite similar to, Casey’s, so I would…
136 00:11:26.770 ⇒ 00:11:46.130 Mustafa Raja: For example, for the, Google, ads spike for, Insomnia, what I did, I went to the Google Ads and, looked for, the options that I have, on my plate. That was, ads, script, and,
137 00:11:46.200 ⇒ 00:11:51.460 Mustafa Raja: doing it via the, ads API.
138 00:11:51.540 ⇒ 00:11:59.439 Mustafa Raja: And, I just gathered the fine, findings that I, I would find, and, …
139 00:11:59.520 ⇒ 00:12:13.529 Mustafa Raja: there’s a spike template that Samuel worked on. What I do is I give it to the AI, and then, give it my thoughts, and have it generate.
140 00:12:13.740 ⇒ 00:12:19.729 Mustafa Raja: An ocean dog, and then trim it according to, my findings and all.
141 00:12:23.890 ⇒ 00:12:24.480 Uttam Kumaran: Okay.
142 00:12:26.530 ⇒ 00:12:27.330 Casie Aviles: What’d you say? Yeah.
143 00:12:28.310 ⇒ 00:12:40.130 Samuel Roberts: Yeah, I had done something similar. That’s actually how I got the template as well, was, generating it. And then I was actually… I chatted back and forth to get more input when I did the Copilot kit one and some of the other options and stuff, but….
144 00:12:40.840 ⇒ 00:12:45.439 Uttam Kumaran: Okay. Yeah, I… I tend to, also kind of, like.
145 00:12:46.260 ⇒ 00:12:56.540 Uttam Kumaran: write down my thoughts, but sometimes I cheat. Maybe I should do what you said, Casey, and just, like, actually think about it a little bit, and then give it my… give habits… fill in the gaps. I usually do the opposite.
146 00:12:56.750 ⇒ 00:13:02.669 Uttam Kumaran: Like, I have it give me a couple ideas, and then I pick one or two, and then I expand on it.
147 00:13:03.350 ⇒ 00:13:09.380 Uttam Kumaran: … Because my main concern is, like, I missed something I’m not considering, but…
148 00:13:09.560 ⇒ 00:13:12.589 Uttam Kumaran: I… maybe it’s, like, some type of mix of both.
149 00:13:14.340 ⇒ 00:13:28.399 Casie Aviles: Yeah, that also makes sense, like, … yeah, sometimes I also have AI suggest, like, you know, potential ways I could tackle something, because, yeah, I agree, sometimes I don’t… I don’t always know everything, like, all the possible options.
150 00:13:28.610 ⇒ 00:13:32.879 Casie Aviles: But yeah, I try to ground it as much as I can with, you know, what I actually observed.
151 00:13:34.010 ⇒ 00:13:34.990 Uttam Kumaran: Yeah….
152 00:13:37.750 ⇒ 00:13:38.540 Casie Aviles: Okay.
153 00:13:39.410 ⇒ 00:13:42.430 Casie Aviles: Alright, so I’ll move on to AI internal…
154 00:13:46.170 ⇒ 00:13:48.230 Casie Aviles: Yeah, I think this one, …
155 00:13:49.050 ⇒ 00:13:52.159 Casie Aviles: Yeah, for the Slack embeddings, I haven’t had, like.
156 00:13:52.830 ⇒ 00:13:58.029 Casie Aviles: movement here, although I did create, like, a loom to help with,
157 00:13:58.730 ⇒ 00:14:03.719 Casie Aviles: I believe, yeah, for Sam, I also created, like, a loom, so he also knows, like.
158 00:14:04.080 ⇒ 00:14:11.200 Casie Aviles: What’s the goal? Like, what’s the… what’s the problem we’re trying to, … Address here.
159 00:14:11.560 ⇒ 00:14:14.260 Casie Aviles: Yeah, and … Yeah, like.
160 00:14:14.260 ⇒ 00:14:22.639 Samuel Roberts: Yeah, that was good, that was helpful. I mean, I didn’t get a chance to do anything with it, but I was able to actually watch it, and that was definitely helpful to get my bearings a little bit, so I appreciate that.
161 00:14:23.070 ⇒ 00:14:24.350 Casie Aviles: Okay, yeah.
162 00:14:24.900 ⇒ 00:14:29.989 Casie Aviles: It’s mostly on Dogster, so there’s a lot of Dogster work there, but yeah.
163 00:14:31.100 ⇒ 00:14:34.720 Casie Aviles: … Okay, so for the other ones….
164 00:14:35.340 ⇒ 00:14:39.960 Uttam Kumaran: Yeah, one thing on the windmill and Daxter failures, I think we should just take a…
165 00:14:40.180 ⇒ 00:14:42.630 Uttam Kumaran: Do, like, a couple-hour session on those.
166 00:14:42.870 ⇒ 00:14:43.810 Uttam Kumaran: …
167 00:14:44.780 ⇒ 00:14:50.949 Uttam Kumaran: I don’t know, I feel like there’s been a lot of errors, so I just want to, like, hit that hard. I don’t know what you think, Casey.
168 00:14:51.240 ⇒ 00:14:55.300 Casie Aviles: Yeah, for Dagster, there are a lot of, errors there.
169 00:14:55.820 ⇒ 00:14:58.430 Casie Aviles: For windmill, those are just mostly…
170 00:14:58.900 ⇒ 00:15:04.579 Casie Aviles: one or two runs that fail, so I just retry them.
171 00:15:07.630 ⇒ 00:15:12.389 Casie Aviles: what else? Yeah, for Dagster, some of… most of those are actually some scripts that…
172 00:15:12.760 ⇒ 00:15:15.679 Casie Aviles: I, yeah, Miguel worked on, and …
173 00:15:15.870 ⇒ 00:15:20.119 Casie Aviles: I haven’t been able to take a look at those extensively.
174 00:15:20.510 ⇒ 00:15:24.300 Casie Aviles: So… And then there’s also, like, ABC…
175 00:15:25.120 ⇒ 00:15:28.100 Casie Aviles: I think the transcript job was also not working.
176 00:15:28.830 ⇒ 00:15:29.600 Casie Aviles: Hmm.
177 00:15:31.140 ⇒ 00:15:32.070 Casie Aviles: Yeah, so….
178 00:15:32.070 ⇒ 00:15:32.560 Samuel Roberts: running on….
179 00:15:32.560 ⇒ 00:15:33.120 Casie Aviles: B.
180 00:15:33.650 ⇒ 00:15:34.599 Samuel Roberts: Oh, sorry, go ahead.
181 00:15:35.150 ⇒ 00:15:37.570 Casie Aviles: Yeah, yeah, I think it would be helpful to…
182 00:15:37.870 ⇒ 00:15:41.639 Casie Aviles: Just dedicate some time to work on all of those.
183 00:15:42.260 ⇒ 00:15:44.590 Casie Aviles: Because they tend to get pushed back.
184 00:15:44.870 ⇒ 00:15:45.400 Uttam Kumaran: Yeah.
185 00:15:48.010 ⇒ 00:15:54.780 Uttam Kumaran: What’s running on Windmill versus Dagster? I saw some of the stuff for the stock embeddings and everything from the loom, but I’m not too familiar with what’s on Windmill.
186 00:15:54.780 ⇒ 00:15:57.050 Casie Aviles: Oh, for windmill, it’s ju- it’s…
187 00:15:57.240 ⇒ 00:16:09.219 Casie Aviles: mainly handling the Zoom stuff, so it’s moving… whenever, like, a Zoom meeting ends, you know, or a cloud recording is available, it gets, …
188 00:16:09.370 ⇒ 00:16:13.089 Casie Aviles: Basically, a payload gets sent to our… to windmill, and.
189 00:16:13.090 ⇒ 00:16:13.860 Samuel Roberts: That’s right.
190 00:16:13.860 ⇒ 00:16:19.570 Casie Aviles: Yeah, it kicks off, like, the export, or, like, the uploading to S3, and then Superbase, yeah.
191 00:16:19.570 ⇒ 00:16:21.809 Samuel Roberts: Got it. Cool. Thanks.
192 00:16:23.080 ⇒ 00:16:26.970 Casie Aviles: But yeah, I would say, windmill, we… there are very, …
193 00:16:27.150 ⇒ 00:16:32.499 Casie Aviles: there’s way less errors that we get there compared to Dungster. We have a lot on downstairs, so…
194 00:16:35.690 ⇒ 00:16:43.930 Casie Aviles: Yeah, I think that’s about it for internal. We have some to-dos here. I mean, we haven’t gotten the chance, I believe, to….
195 00:16:44.630 ⇒ 00:16:45.370 Samuel Roberts: Yeah.
196 00:16:45.690 ⇒ 00:16:48.489 Casie Aviles: Move forward with most of these, so….
197 00:16:48.990 ⇒ 00:16:49.710 Uttam Kumaran: Okay.
198 00:16:52.600 ⇒ 00:17:01.049 Casie Aviles: Yeah, data platform… I haven’t… there’s, like, just one thing I have to check here, which is Glockify, I think that’s what Amber raised.
199 00:17:02.050 ⇒ 00:17:02.900 Casie Aviles: ….
200 00:17:02.900 ⇒ 00:17:07.740 Uttam Kumaran: Yeah, that’s… that’s all okay. I think… and then, how’s it going with Insomnia stuff?
201 00:17:09.089 ⇒ 00:17:17.689 Casie Aviles: Yeah, for insomnia, I also created, like, a spike for them, so I’ve been… I’ve already, like, automated
202 00:17:17.819 ⇒ 00:17:23.969 Casie Aviles: … their campaign, so for Braze, so that’s already automated.
203 00:17:24.259 ⇒ 00:17:30.219 Casie Aviles: There’s just one fix that… Robert asked me to do, which is just regarding the
204 00:17:30.689 ⇒ 00:17:39.279 Casie Aviles: deployed date, but otherwise, it’s working… I think it’s working well. I have it on a separate sheet here, and it allows me to just
205 00:17:39.449 ⇒ 00:17:46.409 Casie Aviles: Because before, we would typically… The process here is we would go to each campaign.
206 00:17:46.779 ⇒ 00:17:54.469 Casie Aviles: You know, click individually, and then we’d have to wait for that, and then get, like, all these data here.
207 00:17:54.729 ⇒ 00:17:57.419 Casie Aviles: And we have to copy-paste that, so that’s very long.
208 00:17:57.689 ⇒ 00:18:04.879 Casie Aviles: It’s a long process. So now, using the API, I’m able to just pull this… just remove the filters.
209 00:18:05.039 ⇒ 00:18:07.069 Casie Aviles: So I’m just able to get this.
210 00:18:08.009 ⇒ 00:18:08.979 Casie Aviles: And I would just…
211 00:18:09.449 ⇒ 00:18:16.419 Casie Aviles: double-check some of that, like, I will just do spot checking, and then I would just copy it into their
212 00:18:16.949 ⇒ 00:18:21.319 Casie Aviles: Into the actual tracker that they use. So that saves me a lot of time.
213 00:18:22.729 ⇒ 00:18:31.919 Casie Aviles: And… but also for Google Ads, I’ve also done Google Ads, and this is what was, you know, this via Polyatomic.
214 00:18:33.109 ⇒ 00:18:35.089 Samuel Roberts: So this is also helpful.
215 00:18:35.349 ⇒ 00:18:36.029 Casie Aviles: …
216 00:18:36.289 ⇒ 00:18:44.179 Casie Aviles: So I did a spike on… for two more things that I needed to automate, which is for DoorDash and ads, Uber ads.
217 00:18:45.089 ⇒ 00:18:52.509 Casie Aviles: So… for that, I was using, like, a browser extension, And…
218 00:18:53.199 ⇒ 00:18:57.839 Casie Aviles: Yeah, it’s like, an automated… what do you call this?
219 00:18:57.969 ⇒ 00:19:00.839 Casie Aviles: browser… Automation that lets us grade.
220 00:19:02.210 ⇒ 00:19:11.550 Casie Aviles: It’s called Axiom, but it’s, yeah, it’s a… it’s a different tool, but maybe I should also… I should… I think I should also explore browser-based for this.
221 00:19:12.330 ⇒ 00:19:17.779 Casie Aviles: But this one just… Running this… I was able to get…
222 00:19:18.010 ⇒ 00:19:22.650 Casie Aviles: At least this, these data, which we need, are mostly just the
223 00:19:23.620 ⇒ 00:19:27.750 Casie Aviles: The cost and the revenue. We need those, …
224 00:19:27.880 ⇒ 00:19:31.810 Casie Aviles: Although it’s not… it’s pretty rough yet. It’s still pretty rough for now.
225 00:19:32.570 ⇒ 00:19:35.799 Casie Aviles: As this is still… just a spike, but…
226 00:19:36.410 ⇒ 00:19:38.810 Casie Aviles: I was able to at least get the data.
227 00:19:40.130 ⇒ 00:19:46.499 Casie Aviles: But yeah, I’m also considering browser-based, so I might have to look into that yet. I haven’t really used browser-based, but…
228 00:19:47.700 ⇒ 00:19:52.150 Casie Aviles: Okay. I think… Yeah, that… before I can decide, like, which…
229 00:19:54.050 ⇒ 00:19:58.880 Casie Aviles: Yeah, I’ll try that first. But yeah, I think that’s… that’s about it for Insomnia.
230 00:20:01.200 ⇒ 00:20:04.679 Casie Aviles: Yeah, so I’ll just, complete.
231 00:20:05.170 ⇒ 00:20:08.800 Casie Aviles: Despite there, and then I can move to implementation.
232 00:20:11.190 ⇒ 00:20:11.880 Casie Aviles: Bye.
233 00:20:12.940 ⇒ 00:20:13.520 Uttam Kumaran: Okay.
234 00:20:16.670 ⇒ 00:20:20.880 Uttam Kumaran: And then, how’s it going for, Interlude?
235 00:20:25.460 ⇒ 00:20:26.430 Uttam Kumaran: Mmm, softball.
236 00:20:26.430 ⇒ 00:20:27.070 Samuel Roberts: Huh?
237 00:20:30.590 ⇒ 00:20:31.320 Samuel Roberts: Nope.
238 00:20:31.320 ⇒ 00:20:36.530 Mustafa Raja: Sorry, I was muted. Can you open this Intel 17 ticket?
239 00:20:39.830 ⇒ 00:20:41.380 Mustafa Raja: Yeah, and scroll down.
240 00:20:43.140 ⇒ 00:20:50.610 Mustafa Raja: Yeah, so, so these, these, the, so this is the review that, that you gave on,
241 00:20:50.800 ⇒ 00:20:53.279 Mustafa Raja: the Notion integration. …
242 00:20:53.930 ⇒ 00:21:02.830 Mustafa Raja: So, out of these 5, 3 of them, were done, and for the last one, I did submit a PR.
243 00:21:03.610 ⇒ 00:21:11.979 Mustafa Raja: once it’s approved, I’d be able to, add LLMSH results with the,
244 00:21:12.110 ⇒ 00:21:14.550 Mustafa Raja: With the… with each iteration.
245 00:21:14.970 ⇒ 00:21:17.479 Mustafa Raja: From DEXAgent.
246 00:21:17.690 ⇒ 00:21:20.309 Mustafa Raja: And for the first one, I believe.
247 00:21:20.930 ⇒ 00:21:25.470 Mustafa Raja: I would need your input on that.
248 00:21:26.520 ⇒ 00:21:27.210 Uttam Kumaran: Okay.
249 00:21:30.940 ⇒ 00:21:34.049 Mustafa Raja: And for today, I’m going to work on the Slack… Slack thing.
250 00:21:34.050 ⇒ 00:21:34.750 Samuel Roberts: boot camp.
251 00:21:36.550 ⇒ 00:21:37.160 Uttam Kumaran: Okay.
252 00:21:44.880 ⇒ 00:21:46.819 Mustafa Raja: I guess VIN is default.
253 00:21:49.170 ⇒ 00:21:49.989 Samuel Roberts: Hell yeah.
254 00:21:51.570 ⇒ 00:21:58.460 Mustafa Raja: Yeah, for the default… there’s a ticket in progress.
255 00:22:02.070 ⇒ 00:22:10.369 Mustafa Raja: Yes, right. for this, I successfully ran tests on…
256 00:22:10.880 ⇒ 00:22:13.599 Mustafa Raja: Let me tell you which websites.
257 00:22:16.730 ⇒ 00:22:32.310 Mustafa Raja: Yeah, on BrainTrust, Vellum Wrap, Arise, and then Lev gave us some of the websites. Out of those, I tested on, HiveMQ and PlanHat, and those successfully ran.
258 00:22:35.930 ⇒ 00:22:40.370 Mustafa Raja: Meaning Diggy was the… Scheduler, and logos.
259 00:22:41.070 ⇒ 00:22:47.729 Mustafa Raja: These scripts are updated, though. A lot of things have been changed in the script.
260 00:22:59.140 ⇒ 00:22:59.850 Casie Aviles: Okay.
261 00:22:59.990 ⇒ 00:23:00.560 Casie Aviles: Bye.
262 00:23:00.560 ⇒ 00:23:01.090 Uttam Kumaran: Okay.
263 00:23:03.010 ⇒ 00:23:07.579 Uttam Kumaran: So, I guess, like, what do you… like, do you think we can demo something today?
264 00:23:08.770 ⇒ 00:23:19.819 Mustafa Raja: We should be able to. Although, I would like, I would like… maybe I can show Sam how it works after this meeting.
265 00:23:24.150 ⇒ 00:23:26.410 Mustafa Raja: Just to get, a review.
266 00:23:27.070 ⇒ 00:23:27.750 Samuel Roberts: Sure.
267 00:23:27.920 ⇒ 00:23:28.350 Mustafa Raja: on this.
268 00:23:36.490 ⇒ 00:23:37.130 Samuel Roberts: Okay.
269 00:23:41.560 ⇒ 00:23:43.699 Samuel Roberts: You have those running locally, is that what it was?
270 00:23:43.700 ⇒ 00:23:46.209 Mustafa Raja: Yeah, for now it’s local.
271 00:23:46.450 ⇒ 00:23:50.149 Mustafa Raja: I believe once we… once we are okay with it.
272 00:23:50.330 ⇒ 00:23:53.929 Mustafa Raja: We can deploy it on cloud and have it run there.
273 00:23:54.530 ⇒ 00:23:55.600 Samuel Roberts: Cool, okay.
274 00:24:00.340 ⇒ 00:24:00.680 Mustafa Raja: And then….
275 00:24:00.680 ⇒ 00:24:01.110 Samuel Roberts: Maybe?
276 00:24:01.110 ⇒ 00:24:13.090 Mustafa Raja: That’s it for this. Once it’s… once it’s, once it’s okay to run, we’d be able to, proceed with some of the signals.
277 00:24:17.680 ⇒ 00:24:21.750 Uttam Kumaran: Okay. Yeah, it would be great to try to knock this out today.
278 00:24:22.130 ⇒ 00:24:29.095 Mustafa Raja: Yeah, although I believe that there are some improvements that we can make in the script, Oh…
279 00:24:29.650 ⇒ 00:24:36.330 Mustafa Raja: But I feel it’s going to be… Improving… as we go.
280 00:24:38.420 ⇒ 00:24:39.090 Uttam Kumaran: Okay.
281 00:24:39.630 ⇒ 00:24:41.740 Uttam Kumaran: Yeah, that’s fine, as long as it’s, like.
282 00:24:42.070 ⇒ 00:24:46.239 Uttam Kumaran: majority of the time, it’s working, that’s what I’m most interested in. We can improve it as we go.
283 00:24:46.240 ⇒ 00:24:47.620 Mustafa Raja: Yeah.
284 00:24:47.990 ⇒ 00:25:07.259 Mustafa Raja: It’s just that, some of the times, it’s a scheduler or a form that it’s not seen before, so I have to look into it and then tell it, okay, this is the… this is the form, this is the scheduler, this is the type of the scheduler. If you see it, mark it as this.
285 00:25:07.490 ⇒ 00:25:13.769 Uttam Kumaran: That’s fine. Yeah, and I believe… I think we’ll build… we’ll build up those cases as we go, but yeah.
286 00:25:13.770 ⇒ 00:25:14.140 Samuel Roberts: Yeah.
287 00:25:14.140 ⇒ 00:25:19.349 Mustafa Raja: It’s going to be our small database of schedulers and forms, I believe.
288 00:25:19.350 ⇒ 00:25:20.040 Uttam Kumaran: Yeah.
289 00:25:20.960 ⇒ 00:25:24.819 Mustafa Raja: To my findings, a lot of the people are using HubSpot.
290 00:25:24.950 ⇒ 00:25:26.809 Mustafa Raja: For meetings and scheduling.
291 00:25:27.490 ⇒ 00:25:28.110 Uttam Kumaran: Okay.
292 00:25:31.010 ⇒ 00:25:33.929 Mustafa Raja: And some of them aren’t using any of those at all.
293 00:25:34.680 ⇒ 00:25:35.070 Samuel Roberts: Hmm.
294 00:25:35.240 ⇒ 00:25:37.019 Mustafa Raja: Don’t skate today.
295 00:25:37.630 ⇒ 00:25:39.010 Mustafa Raja: The phone was custom.
296 00:25:43.290 ⇒ 00:25:43.860 Uttam Kumaran: Okay.
297 00:25:46.490 ⇒ 00:25:50.650 Mustafa Raja: Yeah, let me know, Sam, if you have time after this meeting, or whenever you have time after this meeting.
298 00:25:51.400 ⇒ 00:25:53.439 Samuel Roberts: Okay, yeah, I’ll ping you afterwards.
299 00:25:56.240 ⇒ 00:25:57.270 Uttam Kumaran: Okay, great.
300 00:25:57.270 ⇒ 00:26:01.779 Mustafa Raja: And for interlude, let me know what my priorities should be.
301 00:26:02.930 ⇒ 00:26:07.340 Uttam Kumaran: Yeah, I’m gonna, … I’m gonna look through and get those prompts.
302 00:26:07.530 ⇒ 00:26:11.960 Uttam Kumaran: … I would like to… …
303 00:26:12.120 ⇒ 00:26:14.350 Uttam Kumaran: Take a look at some outputs today.
304 00:26:14.610 ⇒ 00:26:17.669 Uttam Kumaran: But yeah, if we can get the basic Slack app going.
305 00:26:18.300 ⇒ 00:26:19.769 Uttam Kumaran: I’m kind of happy with that.
306 00:26:19.770 ⇒ 00:26:22.089 Mustafa Raja: For the first generation, right?
307 00:26:22.660 ⇒ 00:26:23.310 Uttam Kumaran: Yeah.
308 00:26:23.310 ⇒ 00:26:29.689 Mustafa Raja: Yeah, and then, yeah, yeah, yeah, let’s, let’s work on that, and hopefully we’ll get it out soon.
309 00:26:30.860 ⇒ 00:26:31.490 Uttam Kumaran: Okay.
310 00:26:32.350 ⇒ 00:26:32.880 Mustafa Raja: Thank you.
311 00:26:37.380 ⇒ 00:26:39.750 Uttam Kumaran: Okay, cool. …
312 00:26:40.000 ⇒ 00:26:48.009 Uttam Kumaran: Yeah, I just… if we don’t have anything else, I just wanted to talk about the session tomorrow. I was thinking maybe we should just walk through…
313 00:26:48.610 ⇒ 00:26:53.979 Uttam Kumaran: like, I can walk through a couple of examples of, you know, Like, a couple of…
314 00:26:54.350 ⇒ 00:27:04.029 Uttam Kumaran: GitHub, like, a couple of, ChatGBT prompts that I created this week, and sort of how I created them, I think that could be helpful.
315 00:27:04.450 ⇒ 00:27:07.679 Uttam Kumaran: I also think one thing that I want to try to…
316 00:27:07.820 ⇒ 00:27:10.760 Uttam Kumaran: Build up. It’s just a dialogue between…
317 00:27:10.910 ⇒ 00:27:24.059 Uttam Kumaran: this team and the rest of the company. So I was also gonna ask ChatGBT to help me think about how do we build, like, how do we make this meeting, like, really, really effective for everyone?
318 00:27:24.490 ⇒ 00:27:25.270 Samuel Roberts: Mmm.
319 00:27:25.270 ⇒ 00:27:33.499 Uttam Kumaran: Partly is that our team is… is… there’s some vocal people, like me, and then there’s also a lot of, like, folks that are a little bit just, like.
320 00:27:33.590 ⇒ 00:27:51.779 Uttam Kumaran: you know, don’t speak up a lot in meetings, and so I want to create an environment where folks can add notes, and so maybe, like, a… I think the one thing that’s been really, really great for everybody, I think, that I’ve noticed is just doing, like, a FigJam retro-style meeting.
321 00:27:52.090 ⇒ 00:27:57.349 Uttam Kumaran: Where maybe we just… The first 10 minutes of the meeting is, like.
322 00:27:57.450 ⇒ 00:28:01.269 Uttam Kumaran: We do a… we do, like, a brainstorm of, like, hey, put in…
323 00:28:01.440 ⇒ 00:28:06.820 Uttam Kumaran: like, I know this is a guy I gotta ask, work with ChatGBT, but maybe it’s, like, put in, like.
324 00:28:06.960 ⇒ 00:28:19.379 Uttam Kumaran: The things that you do every day, put in things that are most painful, put in things that you think you can automate, and then we can then take another step to, like.
325 00:28:19.740 ⇒ 00:28:24.090 Uttam Kumaran: Go vote on… on things that… We want to discuss.
326 00:28:24.200 ⇒ 00:28:30.939 Uttam Kumaran: And then maybe we chat about, like, the ways that the team can actually go automate those things, whether it’s…
327 00:28:31.070 ⇒ 00:28:46.199 Uttam Kumaran: creating a prompt, whether it’s something that already exists, whether it’s something that we can do in, you know, pretty quickly, or whether it’s something long-term that we have to notice. So partly it’s, like, a farm for ideas, partly it’s, like, training.
328 00:28:46.680 ⇒ 00:28:53.739 Uttam Kumaran: I don’t know. I feel like, out of everything we’ve done, the sort of sticky note fig jam exercises have always been…
329 00:28:53.910 ⇒ 00:28:56.110 Uttam Kumaran: Really, really productive, so….
330 00:28:58.760 ⇒ 00:29:01.280 Samuel Roberts: Yeah, I mean, if they’ve worked in the past, I think it’s probably…
331 00:29:01.830 ⇒ 00:29:04.570 Samuel Roberts: You know, I don’t think I’ve seen that yet, but yeah.
332 00:29:04.790 ⇒ 00:29:06.989 Samuel Roberts: … That sounds cool.
333 00:29:08.770 ⇒ 00:29:10.639 Casie Aviles: That sounds so good.
334 00:29:11.420 ⇒ 00:29:12.230 Uttam Kumaran: Okay.
335 00:29:15.310 ⇒ 00:29:28.529 Uttam Kumaran: I think partly, it’s, like, I want to get ideas, but also, like, I’m more interested in, like, getting folks to start using the prompt creators, prompt optimizers, and so I can also walk through some examples.
336 00:29:28.620 ⇒ 00:29:36.039 Uttam Kumaran: But I’m also interested to see that people are thinking about optimization and automation every week.
337 00:29:36.430 ⇒ 00:29:42.309 Uttam Kumaran: So, let me put together an agenda for that, and I can lead that, and then we can kind of see where it goes.
338 00:29:43.640 ⇒ 00:29:44.690 Samuel Roberts: Okay.
339 00:29:45.490 ⇒ 00:29:46.330 Samuel Roberts: That sounds cool.
340 00:29:47.130 ⇒ 00:29:47.790 Uttam Kumaran: Okay.
341 00:29:51.600 ⇒ 00:29:52.130 Uttam Kumaran: Okay.
342 00:29:52.130 ⇒ 00:29:52.450 Samuel Roberts: Cool.
343 00:29:52.450 ⇒ 00:29:54.380 Uttam Kumaran: Great. Anything else?
344 00:29:56.680 ⇒ 00:29:57.950 Samuel Roberts: Nope, that I have.
345 00:29:58.990 ⇒ 00:30:00.459 Casie Aviles: I think that’s all from me.
346 00:30:00.460 ⇒ 00:30:01.190 Samuel Roberts: Yeah.
347 00:30:01.940 ⇒ 00:30:02.570 Uttam Kumaran: Okay.
348 00:30:03.320 ⇒ 00:30:05.940 Uttam Kumaran: Alright, perfect. Thank you guys, I’ll chat on Slack.
349 00:30:06.170 ⇒ 00:30:06.580 Mustafa Raja: Thank you.
350 00:30:06.580 ⇒ 00:30:07.370 Samuel Roberts: Alrighty.
351 00:30:07.710 ⇒ 00:30:08.430 Uttam Kumaran: Okay.
352 00:30:08.430 ⇒ 00:30:09.370 Mustafa Raja: Thank you. Bye.