Videos Byv311hdoHE

The Dirty Secret of Forward Deployed Engineering — Natalie Meurer, Sierra

index_state ready data_status ok

AI Engineer· published 2026-07-28· 0:16:49· en-US· indexed 2026-08-10 19:38

Open on YouTube

41 shot(s).

keyframes kept every frame deduplicated

What was stored

cues: 174; whisperx 174
chunks: 30; from 174 cues
keyframes: 28; kept of 41 captured
frames with text: 28; 414 lines read
chapters: 10; from the source metadata
keyframe bytes: 5.9 MB; word timings on 174 cues

Provenance

Each pipeline stage, its state and the model that produced it
stage	state	model	started	took
`fetch`	done	—	2026-08-09 22:14	0s
`stt`	done	—	2026-08-09 05:50	19s
`chunk`	done	—	2026-08-09 05:51	0s
`text_embed`	done	—	2026-08-10 19:38	1s
`keyframe`	done	—	2026-08-09 05:51	2m 08s
`ocr`	done	—	2026-08-09 05:53	9s
`frame_embed`	done	—	2026-08-10 19:38	5s

Frames, and what the machine read

0:02 #0 done2 line(s)

shot 0·sharpness 454.1
1. AlEngineer0.95
2. World's Fair0.97
0:03 #1 done2 line(s)

shot 1·sharpness 665.2
1. AIEngineer0.95
2. World's Fair0.99
0:10 #2 done24 line(s)

shot 2·sharpness 2746.2
1. LAB & PLATINUM SPONSORS0.99
2. Amazon AGI Lab0.98
3. ANTHROP\C1.00
4. Google DeepMind1.00
5. MINIMAX0.94
6. OpenAI0.92
7. Akamai1.00
8. arize0.93
9. aws1.00
10. Braintrust bright data0.98
11. B1.00
12. Browserbase1.00
13. docker1.00
14. :neo4j0.92
15. ORACLE1.00
16. PayPal1.00
17. qodo1.00
18. reducto1.00
19. Sonar1.00
20. Makers of0.99
21. togetherai1.00
22. Unblocked1.00
23. WorkOS1.00
24. SonarQube1.00
0:25 #3 done2 line(s)

shot 3·sharpness 201.0
1. AlEngineer0.99
2. World's Fair1.00
0:52 #4 done12 line(s)

shot 4·sharpness 1542.2
1. AlEngineer0.99
2. World's Fair0.97
3. The Dirty Secret of1.00
4. Forward Deployed0.97
5. PRESENTED BY1.00
6. Engineering1.00
7. Microsoft1.00
8. Natalie Meurer1.00
9. Head of Agent Engineering,1.00
10. Sierra1.00
11. Engineering the future of Al0.97
12. World's Fair0.98
1:20 #5 done13 line(s)

shot 5·sharpness 1680.8
1. AlEngineer0.98
2. World'sFair1.00
3. Georgetown1.00
4. 2012 - 20160.97
5. Palantir1.00
6. 2016-20211.00
7. PRESENTED BY1.00
8. Microsoft1.00
9. Head of Agent Engineering, Sierra1.00
10. Nat Meurer1.00
11. TRACK 8· JUNE 30, 20260.96
12. World's Fair0.97
13. Forward Deployed Engineering0.99
1:48 #6 skipped

shot 6·duplicate of #5
2:26 #7 done16 line(s)

shot 7·sharpness 1733.0
1. AlEngineer0.98
2. The Sierra blog + Met the Al agent engineer0.90
3. World'sFair0.98
4. Meet the Al agent engineer1.00
5. Natale Meurer0.98
6. Software Engineer0.99
7. ×00.59
8. Aly 11, 20240.74
9. Al Engineer0.99
10. Agent Engineer0.99
11. Head of Agent Engineering, Sierra1.00
12. https://sierra.ai/blog/meet-the-ai-agent-engineer1.00
13. Nat Meurer1.00
14. TRACK 8· JUNE 30, 20260.96
15. World's Fair0.96
16. Forward Deployed Engineering1.00
2:50 #8 done6 line(s)

shot 8·sharpness 1013.2
1. AlEngineer0.97
2. World'sFair0.97
3. The dirty secret1.00
4. TRACK 8· JUNE 30, 20260.96
5. World's Fair0.97
6. Forward Deployed Engineering1.00
2:56 #9 skipped

shot 9·duplicate of #8
4:06 #10 done25 line(s)

shot 10·sharpness 1503.9
1. Palantir Job Postings, c. 20161.00
2. AlEngineer0.99
3. World's Fair0.96
4. ENGINEERING1.00
5. Forward Deployed Software Engineer1.00
6. Palo Alto, CA0.99
7. New York, NY1.00
8. Washington, D.C.1.00
9. London, United Kingdom1.00
10. Ottawa, Canada1.00
11. Paris, France1.00
12. Canberra, Australia0.99
13. Helsinki, Finland1.00
14. Stockholm, Sweden0.99
15. Seattle, WA0.98
16. Sydney, Australia1.00
17. Oslo, Norway1.00
18. Copenhagen, Denmark1.00
19. Melbourne, Australia0.99
20. Tel Aviv, Israel1.00
21. Tokyo, Japan1.00
22. Source: palantir.com (the Wayback Machine - 2016)0.99
23. TRACK 8· JUNE 30, 20260.96
24. World's Fair0.99
25. Forward Deployed Engineering0.98
4:31 #11 done8 line(s)

shot 11·sharpness 1476.9
1. AlEngineer0.99
2. World'sFair0.97
3. How most people think1.00
4. of forward deployed0.99
5. engineering.1.00
6. TRACK 8· JUNE 30, 20260.96
7. World'sFair1.00
8. Forward Deployed Engineering0.99
4:41 #12 done9 line(s)

shot 12·sharpness 1261.2
1. AlEngineer0.97
2. Forward Deployed Engineering: A Timeline1.00
3. World'sFair0.96
4. Q0.91
5. 20081.00
6. Platform stability1.00
7. TRACK 8· JUNE 30, 20260.96
8. World'sFair1.00
9. Forward Deployed Engineering1.00
4:48 #13 done22 line(s)

shot 13·sharpness 2005.9
1. AlEngineer0.98
2. World'sFair1.00
3. scp release-agent-2026.06.29-linux-amd64.tar.gz \0.99
4. fde@atlas-prod-us-east-1a-03:/var/tmp/1.00
5. 40.98
6. # Connect to the target host0.98
7. ssh fde@atlas-prod-us-east-1a-031.00
8. What forward0.99
9. # Install the release0.99
10. cd /var/tmp1.00
11. deployed engineering1.00
12. tar -xzf1.00
13. actually was.1.00
14. release-agent-2026.06.29-linux-amd64.tar.gz1.00
15. cd release-agent-2026.06.291.00
16. chmod +x init.sh0.98
17. sudo ./init.sh --env=prod0.98
18. # Verify the service1.00
19. systemctl status agentd1.00
20. TRACK 8· JUNE 30, 20260.94
21. World's Fair0.99
22. Forward Deployed Engineering1.00
5:21 #14 done16 line(s)

shot 14·sharpness 1586.6
1. AlEngineer0.97
2. World'sFair1.00
3. From:1.00
4. [email protected]0.99
5. What forward1.00
6. To:1.00
7. you1.00
8. Subject: Paul unplugged the Palantir instance again0.99
9. deployed engineering1.00
10. Hi Nat,0.95
11. actually was.0.99
12. Paul unplugged the server rack again. Sorry about that. Could you investigate ASAP?0.99
13. Customer1.00
14. TRACK 8· JUNE 30,20260.95
15. World'sFair1.00
16. Forward Deployed Engineering1.00
5:48 #15 done13 line(s)

shot 15·sharpness 1475.9
1. AlEngineer0.98
2. Forward Deployed Engineering: A Timeline0.99
3. World'sFair1.00
4. Q0.74
5. PRESENTED BY1.00
6. Microsoft1.00
7. 20081.00
8. 20121.00
9. Platform stability1.00
10. Data integration1.00
11. TRACK 8· JUNE 30, 20260.96
12. World'sFair0.99
13. Forward Deployed Engineering1.00
5:58 #16 done10 line(s)

shot 16·sharpness 1802.6
1. AlEngineer0.99
2. World'sFair1.00
3. RIVIERA1.00
4. PRESENTED BY1.00
5. NOW PLAYING:1.00
6. Microsoft1.00
7. NOTHING1.00
8. TRACK 8· JUNE 30, 20260.96
9. World's Fair0.98
10. Forward Deployed Engineering1.00
6:28 #17 done26 line(s)

shot 17·sharpness 3154.2
1. AlEngineer0.99
2. World's Fair0.98
3. PRESENTED BY1.00
4. this data integration1.00
5. software. Think it1.00
6. Hi, check out0.99
7. could be useful?0.97
8. useful, but all my1.00
9. data is in these0.99
10. Yeah, that could be1.00
11. 20 places.0.96
12. useful, I need that data0.95
13. Okay, so for this to be0.99
14. in this one place. That1.00
15. cool?0.98
16. Sure - if you embed0.97
17. with us onsite and1.00
18. do it for us, I'd1.00
19. use it.0.99
20. 0-000D00.61
21. Microsoft1.00
22. Palanti0.92
23. TODO:0.99
24. TRACK 8· JUNE 30, 20260.96
25. World's Fair0.98
26. Forward Deployed Engineering1.00
7:07 #18 skipped

shot 18·duplicate of #12
7:20 #19 done40 line(s)

shot 19·sharpness 1552.0
1. AlEngineer0.99
2. The Era of Slate0.99
3. World's Fair0.98
4. Application Name1.00
5. Page Title1.00
6. Selected tab0.99
7. Another tab Another tab0.97
8. Show Filters0.95
9. Filter21.00
10. Filter40.95
11. valuel0.99
12. value20.99
13. YY-MM-000.92
14. YYY-MM-OD0.87
15. Put Your Title Here0.99
16. Put your subheader here1.00
17. Description1.00
18. Description1.00
19. Description1.00
20. Description1.00
21. Description1.00
22. 4,7901.00
23. 50.76%1.00
24. 35.4k1.00
25. $67.1m1.00
26. -31℃1.00
27. Put Your Title Here0.99
28. Put your subheader here1.00
29. 0.0b0:0.70
30. Explanatory Subheader1.00
31. immediately egible. nclude notes on how the values are0.89
32. Consider adding explanatory text in order to make this chart0.98
33. derived, how you should interpret the figures, or other0.98
34. informaton ust bea im mid if youare using fiiteers,the0.79
35. informatlele e lil ill lel change.0.76
36. Also note, thsis already a lot of text. Try to keep it short and0.91
37. sweet.0.99
38. TRACK 8· JUNE 30, 20260.96
39. World's Fair0.99
40. Forward Deployed Engineering1.00
8:13 #20 skipped

shot 20·duplicate of #19
8:31 #21 skipped

shot 21·duplicate of #12
9:16 #22 done5 line(s)

shot 22·sharpness 2120.5
1. AlEngineer0.99
2. World'sFair1.00
3. TRACK 8· JUNE 30, 20260.96
4. World's Fair0.98
5. Forward Deployed Engineering1.00
9:59 #23 done15 line(s)

shot 23·sharpness 1601.7
1. AlEngineer0.97
2. Total FDE Jobs to be Done0.97
3. World'sFair0.96
4. Enablement1.00
5. Custom1.00
6. solutions1.00
7. Data integration1.00
8. DevOps1.00
9. 20081.00
10. 20121.00
11. 20161.00
12. 20201.00
13. TRACK 8· JUNE 30, 20260.96
14. World'sFair1.00
15. Forward Deployed Engineering1.00

Transcript

174 cues· 2,993 words· 16,784 chars

chunk 0 · 0:12–0:55 · 149 words · 828 chars
0:12 Thanks, guys, for coming.
0:13 I know right after lunch, this is prime time sleepiness.
0:18 I'm curious, just to get a sense for the audience, who here is a forward deployed engineer?
0:22 OK, who here is hiring forward deployed engineers?
0:25 OK, helpful.
0:26 Anyone else in between?
0:27 You can raise your hand, just like if I didn't cover it.
0:29 OK, I see you.
0:31 Double hand raise.
0:33 Well, I'm Natalie Muir.
0:34 I am the head of agent engineering at Sierra.
0:36 I've been at the company about two years.
0:38 I'll tell you a little bit about that.
0:40 But honestly, I'm here to talk about forward-to-point engineering.
chunk 1 · 0:42–1:20 · 135 words · 722 chars
0:42 So we're actually gonna spend a little bit less time talking about Sierra and a lot more time talking about the history of forward-to-point engineering and the dirty secret, I like to call it, of the domain.
0:53 And you all can tell me how controversial it is.
0:56 So just a little bit about me, which will give you a sense for sort of how I'm coming at this domain, at the whole domain of forward-to-point engineering.
1:03 I actually started off as a policy nerd.
1:05 So I was at Georgetown.
1:07 I studied in the School of Foreign Service there, and so I was obsessed with basically tech policy.
chunk 2 · 1:12–1:57 · 151 words · 812 chars
1:12 I learned to code and then swindled my way, or earned my way, depending upon your perspective, into a job at Palantir where I spent 2016 to 2021.
1:22 And there I worked on the privacy team and I was an infrastructure engineer and also a forward deployed engineer with law enforcement and defense primarily.
1:30 So I've seen forward deployed engineering up close and I'm gonna talk a lot about the history of not just my time at Palantir but actually the time that predated me to talk about basically the origins of the entire discipline.
1:41 And then I actually wanted to get a business degree, of all things, so I went to Stanford, and now I'm at Sierra from 24 to 26.
chunk 3 · 1:49–2:31 · 135 words · 731 chars
1:49 And one of the things that I think is quite interesting, and when I started at Sierra, actually a little known fact, is that I joined the deployment team.
1:58 And at the time, I hated the name.
2:00 I thought, this is not what we're doing, A.
2:03 And B, it sort of is a nonsensical name, so to speak.
2:06 And so I actually wrote an article in July of 2024, so about six months after I started, about two years ago now, on this concept of the agent engineer.
2:16 And the idea was that agent engineering was actually a sub-discipline of AI engineering, which we're of course here to talk about.
chunk 4 · 2:23–3:02 · 135 words · 752 chars
2:23 And the vision was that agent engineering had the same level of customer accountability as forward deployed engineering, but was actually a domain unto itself.
2:31 And I think since then, in the past two years, we've actually seen a lot of folks talking about agent engineering as a discipline, and now we're actually seeing harness engineering as sort of a new discipline, sort of as almost a subset of agent engineering.
2:42 And so if you're confused, that's sort of what this talk is about.
2:46 And so the dirty secret of forward-deployed engineering, I won't make you wait until the end, is it doesn't exist.
chunk 5 · 2:53–3:32 · 136 words · 722 chars
2:53 It is a term that is meant to describe so many things that it sort of means nothing at this point.
3:00 But nonetheless, it's sort of the hottest job in AI.
3:02 And so what I hope to convince you of over the next, let's say 15 minutes, is A, that this is true, and B, that this doesn't matter.
3:12 And so I'm gonna dive in, and we're gonna start with tracing the history of forward deployed engineering from Palantir.
3:19 So we're gonna start in 2008, and we're gonna basically come back to the present.
3:22 And then I'll share with you at least my perspective
chunk 6 · 3:25–3:59 · 118 words · 668 chars
3:25 on what forward deployed engineering is today, why you should want to or not want to be a forward deployed engineer, and why we're actually all forward deployed engineers in the end.
3:34 So first off, I'd actually like to start in 2016, and then we'll go a little back in time.
3:39 And I actually pulled basically Palantir job postings in 2016, this is when I started, and what drew me into Palantir, was effectively this concept of a forward deployed software engineer.
3:50 And this was an emphasis on the forward deployed part.
3:53 I mean, they're leading, these days you lead with, you know, you speak a certain language, you have these skill sets, they were leading with location.
chunk 7 · 4:00–4:42 · 113 words · 679 chars
4:00 And there's a reason for that that we'll get to, but forward deployed literally meant sitting with customers, being on the ground, and in part, as we'll talk about, this was because so much of Palantir's ecosystem was on-prem deployments.
4:14 And so when we think of forward-to-point engineering, though, we kind of think of this.
4:18 And so this is sort of the prototypical forward-to-point engineer right out of college parachuting out of a helicopter and, of course, fixing some sort of runtime issue in the air.
chunk 8 · 4:30–5:14 · 130 words · 743 chars
4:30 But in practice, forward deployed engineering actually started as something more akin to DevOps.
4:37 And so platform stability really constituted the bulk of the job early on.

Chapters

0:00 Introduction: the dirty secret of FDE
1:54 FDE as a subdiscipline of AI engineering
3:46 Palantir origins: leading with location
4:51 Firefighting platform stability
6:37 DevOps, data, and the ontology in Slate
9:02 Why FDE is not one thing
10:46 Every company hires a different vintage
12:28 When code gets cheap, outcomes matter
13:55 Pricing: seats vs usage vs outcome
15:26 Agent engineering is FDE reborn