Videos li0SaBt9RDM

The AI Skill I Rely On Daily — Priscila Andre de Oliveira, Sentry

index_state ready data_status ok

AI Engineer· published 2026-05-27· 0:17:04· en-US· indexed 2026-08-10 19:51

Open on YouTube

55 shot(s).

keyframes kept every frame deduplicated

What was stored

cues: 209; whisperx 209
chunks: 30; from 209 cues
keyframes: 36; kept of 55 captured
frames with text: 36; 804 lines read
chapters: 10; from the source metadata
keyframe bytes: 5.4 MB; word timings on 209 cues

Provenance

Each pipeline stage, its state and the model that produced it
stage	state	model	started	took
`fetch`	done	—	2026-08-10 14:27	2m 23s
`stt`	done	—	2026-08-10 14:29	16s
`chunk`	done	—	2026-08-10 14:30	0s
`text_embed`	done	—	2026-08-10 19:51	1s
`keyframe`	done	—	2026-08-10 14:30	1m 31s
`ocr`	done	—	2026-08-10 14:31	20s
`frame_embed`	done	—	2026-08-10 19:51	6s

Frames, and what the machine read

0:03 #0 done2 line(s)

shot 0·sharpness 662.1
1. AlEngineer0.98
2. EUROPE1.00
0:08 #1 done2 line(s)

shot 1·sharpness 814.7
1. PRESENTINGSPONSOR1.00
2. Google DeepMind1.00
0:12 #2 done3 line(s)

shot 2·sharpness 905.6
1. PLATINUM SPONSORS0.98
2. # Braintrust0.96
3. WorkOS OpenAI0.94
0:19 #3 done5 line(s)

shot 3·sharpness 371.3
1. Comprehend First,1.00
2. Code Later1.00
3. :neer0.95
4. Modal0.97
5. LESS.0.95
0:28 #4 done10 line(s)

shot 4·sharpness 1991.9
1. SENTRY0.91
2. Comprehend First,1.00
3. AIE1.00
4. Code Later1.00
5. ★0.99
6. The AI Skill I Rely On Daily0.96
7. Priscila Oliveira1.00
8. Software Engineer @ Sentry0.98
9. 1/231.00
10. Google DeepMind0.98
1:02 #5 done13 line(s)

shot 5·sharpness 2051.5
1. Priscila Oliveira1.00
2. AIE1.00
3. Based in Austria0.99
4. Mom of a 2-year-old toddler1.00
5. ★1.00
6. ★1.00
7. ★1.00
8. Maintainer @ Verdaccio0.99
9. Co-organizer @ ViennaJS0.99
10. Senior Software Engineer @ Sentry0.99
11. 2/231.00
12. Braintrust1.00
13. WorkOS OpenAI0.96
1:41 #6 done9 line(s)

shot 6·sharpness 1863.1
1. Agent Manager @ Sentry0.98
2. *0.70
3. AIE1.00
4. ★1.00
5. ★1.00
6. 4/231.00
7. AlEngineer0.98
8. EUROPE1.00
9. :neer0.89
1:57 #7 done8 line(s)

shot 7·sharpness 1879.3
1. AIE1.00
2. How we ship code is fundamentally0.99
3. changing.1.00
4. ★1.00
5. 6/231.00
6. AlEngineer0.97
7. EUROPE1.00
8. eer1.00
2:30 #8 done20 line(s)

shot 8·sharpness 2564.1
1. Since December 2025, I only prompt.0.99
2. Real PRs, all co-authored with Claude. Reviewed by Cursor,0.99
3. Seer, and humans.1.00
4. AIE1.00
5. fix(agents): Add error boundary around ModelsTable1.00
6. sentry1.00
7. ★1.00
8. ★1.00
9. ★1.00
10. ★0.99
11. feat(conversations): Collapse tool calls in message bubbles1.00
12. sentry1.00
13. fix(langchain): use runName argument to fix unknown_chain spans0.99
14. sentry-javascript1.00
15. feat(browser): Add browser platform support with Playwright1.00
16. testing-ai-sdk-integrations1.00
17. Every commit tagged Co-Authored-By: Claude0.99
18. and many more...1.00
19. 7/231.00
20. Engineering the future of Al0.99
2:44 #9 skipped

shot 9·duplicate of #8
3:10 #10 done49 line(s)

shot 10·sharpness 2126.9
1. This is Sentry.1.00
2. Application monitoring software considered "not bad" by0.98
3. millions of developers.0.99
4. Error monitoring0.98
5. Logging1.00
6. Session replay1.00
7. Automatically capture exceptions1.00
8. and stack traces0.97
9. See logs in context with errors and0.99
10. performance issues0.98
11. Watch real user sessions to see0.99
12. what went wrong1.00
13. ★1.00
14. AIE1.00
15. B Tracing0.88
16. Metrics1.00
17. Profiling0.97
18. Find bottlenecks, broken requests.0.99
19. Track application performance1.00
20. Pinpoint the functions and1.00
21. ★1.00
22. and understand application flow1.00
23. and usage over time with0.99
24. lines of code responsible for1.00
25. ★1.00
26. ★1.00
27. ★1.00
28. end-to-end.1.00
29. custom metrics.1.00
30. performance issues.1.00
31. Agent monitoring1.00
32. Seer: Al Debugging Agent New0.97
33. Track all agent runs, error rates.0.99
34. Catch breaking changes, automatically root cause issues in production, and fix1.00
35. LLM calls, tokens used, and1.00
36. what you missed.1.00
37. tool executions.0.98
38. Requires additional setup0.98
39. 20101.00
40. 15+1.00
41. 100k+1.00
42. -4000.89
43. Founded1.00
44. Years of code1.00
45. Organizations1.00
46. Employees1.00
47. 8/230.95
48. Engineering the future of Al0.97
49. gined0.92
3:51 #11 skipped

shot 11·duplicate of #7
4:05 #12 done35 line(s)

shot 12·sharpness 2338.7
1. Vibe coding is great. And we do it too.1.00
2. All vibe-coded at Sentry. All used by Sentry.0.98
3. 00.75
4. 白0.90
5. ∆0.65
6. AIE1.00
7. Abacus1.00
8. Warden1.00
9. Junior1.00
10. AI SDK0.99
11. ★1.00
12. Al coding tool usage1.00
13. Agent-powered1.00
14. Slack bot that0.99
15. Testing1.00
16. ★1.00
17. ★1.00
18. tracker1.00
19. code review on PRs0.96
20. investigates issues1.00
21. Integration tests1.00
22. getsentry/abacus1.00
23. getsentry/warden1.00
24. getsentry/junior1.00
25. across Al providers1.00
26. Open Source1.00
27. Fair Source1.00
28. Open Source0.97
29. getsentry/testing-1.00
30. ai-sdk-1.00
31. integrations1.00
32. Open Source1.00
33. 9/231.00
34. AlEngineer0.96
35. EUROPE1.00
4:36 #13 done32 line(s)

shot 13·sharpness 2351.6
1. Vibe coding is great. And we do it too.1.00
2. All vibe-coded at Sentry. All used by Sentry.0.99
3. 00.80
4. 白0.91
5. AIE1.00
6. Abacus1.00
7. Warden1.00
8. Junior1.00
9. AI SDK0.99
10. Al coding tool usage1.00
11. Agent-powered1.00
12. Slack bot that0.99
13. Testing1.00
14. ★1.00
15. ★1.00
16. tracker1.00
17. code review on PRs0.97
18. investigates issues1.00
19. Integration tests1.00
20. getsentry/abacus1.00
21. getsentry/warden1.00
22. getsentry/junior1.00
23. across Al providers1.00
24. Open Source1.00
25. Fair Source1.00
26. Open Source1.00
27. getsentry/testing-1.00
28. ai-sdk-1.00
29. integrations1.00
30. Open Source1.00
31. 9/231.00
32. Engineering the future of Al1.00
5:14 #14 done8 line(s)

shot 14·sharpness 379.7
1. Vibe coding is great. And we do it too.0.99
2. Abacus1.00
3. Warden1.00
4. Junior1.00
5. AISDK1.00
6. Testing1.00
7. AIE0.85
8. LESS0.87
5:29 #15 done8 line(s)

shot 15·sharpness 419.2
1. Vibe coding is great. And we do it too.1.00
2. Abacus1.00
3. Warden1.00
4. Junior1.00
5. AISDK0.99
6. Testing1.00
7. AlEngine0.97
8. EUROPE1.00
6:10 #16 skipped

shot 16·duplicate of #12
6:29 #17 skipped

shot 17·duplicate of #13
6:52 #18 skipped

shot 18·duplicate of #8
7:08 #19 skipped

shot 19·duplicate of #8
7:48 #20 skipped

shot 20·duplicate of #8
8:07 #21 skipped

shot 21·duplicate of #8
8:45 #22 done24 line(s)

shot 22·sharpness 2492.1
1. Al made me faster. But not how you think.1.00
2. The speed gain isn't in writing code - it's in understanding0.97
3. it.1.00
4. BEFORE AI0.96
5. AFTER AI1.00
6. ★1.00
7. AIE1.00
8. × Search the codebase to find where a0.98
9. You: why do we not have an error1.00
10. ★1.00
11. function is used0.99
12. boundary there?0.99
13. ★1.00
14. ★1.00
15. x Open multiple PRs to understand a pattern0.98
16. You: where is ModelsTable used?1.00
17. ×Wait for a colleague in another timezone0.98
18. You: how does the conversations table1.00
19. × Piece together git blame, commit by commit0.97
20. work?1.00
21. Same goal – align my mental model with the codebase. Now in minutes, not hours.0.98
22. 14/231.00
23. AlEngineer0.97
24. EUROPE1.00
9:03 #23 skipped

shot 23·duplicate of #22

Transcript

209 cues· 2,060 words· 10,749 chars

chunk 0 · 0:14–0:54 · 96 words · 507 chars
0:14 Hello, everyone.
0:16 Today, I'm going to share with you how I use AI at Sentry and the skill I use the most in my day-to-day work.
0:24 But before we dive into that, let me tell you who I am.
0:29 My name is Priscila.
0:30 I'm a Brazilian based in Vienna, Austria.
0:33 I'm a mom of a two years old, very energetic toddler.
0:37 I'm a maintainer of Verdatio, an open source NPM registry.
0:43 I'm a co-organizer of ViennaJS, a very traditional meetup in Vienna, and we talk all about JavaScript.
chunk 1 · 0:51–1:32 · 88 words · 475 chars
0:51 And I'm a senior software engineer at Sentry.
0:54 Yesterday, someone told me that I don't look like a software engineer, but guess what?
1:00 I am.
1:03 So my official title is senior software engineer, but I have given me a little promotion, and I am now an agent manager.
1:14 No salary raise, but at least my reports, they don't complain.
1:20 Yes, this was me at work a few weeks ago.
chunk 2 · 1:25–2:09 · 79 words · 448 chars
1:25 My colleague Dominique Dorfmeister found it funny to see me managing a couple of agents and took this picture.
1:33 Yeah, luckily I have three monitors, so that works pretty well.
1:41 This is my new reality.
1:43 So this is how I feel, actually, orchestrating a bunch of agents.
1:51 Yeah, it's weird, but it's fun.
chunk 3 · 1:55–2:39 · 76 words · 424 chars
1:55 And the industry is changing, right?
1:58 That's why you all are here.
2:00 And I'm also adapting.
2:03 Since last year, since December 2025, I haven't coded anymore.
2:09 I'm only prompting.
2:11 Yes, and even this presentation was created by a skill.
2:17 Yeah, I didn't do anything.
2:25 So as you can see, these are some of my recent contributions to Sentry.
chunk 4 · 2:30–3:12 · 76 words · 456 chars
2:30 And I created a few PRs together with my favorite teammate, Claude.
2:36 And it's not just bug fixes.
2:39 It's also features, refactors, cross-report stories, contributions.
2:45 So it's real.
2:47 It's working.
2:51 So this is Sentry.
2:54 Maybe you don't know Sentry, but we are very well known for error and performance monitoring.
3:00 But we have grown into a full observability platform.
chunk 5 · 3:04–3:48 · 97 words · 520 chars
3:04 So we have error monitoring.
3:07 We have metrics.
3:08 We have profiling.
3:09 We have also agentic tools like
3:13 to monitor your agent tech platforms.
3:17 Yes, the code base is very complex.
3:20 And it was founded in 2010.
3:23 It has 15 plus years of code.
3:26 We got around 400 employees around the globe.
3:30 We have 100K organizations depending on this code base working every day.
chunk 6 · 3:37–4:21 · 100 words · 511 chars
3:37 And as employee, I also depend on this codebase working because I get my salary from it, right?
3:42 So I don't want to just ship a slop code.
3:46 Yeah, so it's a serious business.
3:51 And we vibe code as well at Sentry.
3:54 Recently, we had a hackathon where we had a few days to just get ourselves familiar with AI and try out new things.

Chapters

0:00 Introduction and speaker background
2:25 Sentry's engineering environment and scale
3:50 AI-driven projects at Sentry
5:48 Maintaining code quality and technical debt
7:35 The role of comprehension in software development
9:38 Analyzing AI usage patterns
10:33 The "Catch Me Up" skill architecture
12:15 Short demo of the "Catch Me Up" skill
13:56 Planning vs. implementation in AI workflows
15:26 Conclusion and key takeaways