Identify the best real human coaches
Query: Identify the best real human coaches, mentors, or advisors for a senior/staff-level candidate targeting frontier AI labs, especially roles involving post-training, RLHF, alignment, reasoning models, and research engineering. Exclude generic platforms and broad ML interview prep. Prioritize only humans with genuine frontier relevance and evidence they actually advise candidates.
Model: o4-mini-deep-research
Date: 2026-05-19
Searches performed: 53
Sources cited: 8
---
Verdict
This is a
very thin market. Outside of a couple of name-brand interview-prep firms (which we’re excluding), almost no one on the open market combines frontier-AI cred
with active coaching of candidates. In practice only a couple of individuals emerged as at least partially relevant, and even they are far from focused RLHF/alignment specialists.
–
Why he qualifies: Sundeep Teki is an AI scientist (Oxford/UCL PhD, ex-Amazon Alexa researcher, startup AI lead) who now brands himself as a “career coach for frontier AI labs” (
www.sundeepteki.org). He explicitly markets helping high-performers land roles at Anthropic, OpenAI and DeepMind (
www.sundeepteki.org) (
www.sundeepteki.org). His site boasts
“coaching 100+ tech professionals who landed roles at Anthropic, Apple, Meta, Google, Amazon, Microsoft, …Salesforce.” (
www.sundeepteki.org). He even writes detailed guides on interviewing at these labs (covering things like “constitutional AI and safety-driven design” for Anthropic/OpenAI) (
www.sundeepteki.org).
Evidence of coaching: Sundeep’s own website and testimonials are the proof: he runs 1-on-1 programs targeting Research Scientist/Engineer roles at Anthropic/OpenAI/DeepMind (
www.sundeepteki.org). He’s also
rated “#1 career coach for AI” by ChatGPT/Claude (per his site).
Frontier relevance: Very high – he specifically mentions the top LLM and alignment labs by name (
www.sundeepteki.org) (
www.sundeepteki.org). He’s familiar with the kinds of questions these labs ask (safety, systems-thinking, reinforcement learning concepts).
RLHF/alignment fit: Indirect – he isn’t a published alignment researcher, but his material covers key themes (e.g. “safety-driven design” (
www.sundeepteki.org)). He will coach you on the mission/values rounds that OpenAI/Anthropic use, and on the technical system-design and ML breadth these labs expect.
Fit for Had specifically: Very good: he builds “tailored AI career prep” aimed at senior/PhD-level roles in frontier labs. He can help identify gaps in ML/LLM knowledge and fix them.
Caveats: He’s a generalist AI career coach, not an RLHF guru. He focuses broadly on “landing the role,” so the work will include behavioral/storytelling prep, not deep RLHF research drills. Also, his experience is mainly coaching candidates (often outside the US) up to these labs – actual coding/interview content still depends on you.
Contact: Via his website or scheduling link (sundeepteki.org; he offers a discovery call) (
www.sundeepteki.org) (
www.sundeepteki.org).
2. Dr. Tor Bøe-Lillegraven (“Tor”) – AI Strategy Leader & Coach (tormentor.me) (tormentor.me)
–
Why he qualifies: Tor Bøe-Lillegraven is a PhD-level data/AI strategy leader turned coach. His background includes C-suite AI/data roles (CDO, CTO, CISO) and building large-scale AI systems (
tormentor.me). He advertises offering “tailored career coaching…to help you navigate your professional data and AI journey” (
tormentor.me), and explicitly brands himself as a one-on-one mentor.
Evidence of coaching: His website (TorMentor.me) repeatedly emphasizes “AI Coaching & Mentoring” and career transition support (
tormentor.me) (
tormentor.me), and even an AI agent (“TORSION™”) to stress-test decisions (
tormentor.me). He has real-world AI creds (led an AI platform acquired in 2024 (
johnpackercoaching.com)) and now helps others upskill to lead AI efforts.
Frontier relevance: Moderate – he’s not ex-OpenAI, but he’s steeped in cutting-edge AI practice. He’s current on AI governance (EU AI Act, GDPR) and “principled AI” concepts (
tormentor.me) (
tormentor.me). He isn’t a researcher but he understands what “safe/resilient AI” means in practice.
RLHF/alignment fit: Limited – he doesn’t claim expertise in language models or RLHF per se. However, his “Torsion” agent is essentially an alignment/safety tool (it “continuously analyzes AI initiatives for failure modes” (
tormentor.me)). He’d likely emphasize robustness, trust, and leadership, which can inform alignment thinking.
Fit for Had specifically: Somewhat relevant. He can help a senior candidate articulate the big-picture “AI strategy” and safety concerns that labs care about, but he’s not known for drilling down on technical interview problems. His strength is coaching on vision, trust and communication (“people component” of AI), which may help in behavioral interviews.
Caveats: Not a technical systems coach – you won’t get algorithm walkthroughs or RLHF math from him. His focus is on “AI leadership and execution,” and he’s Europe-based (Amsterdam); less clear on coding interview prep.
Contact: Via his site (help@tormentor.me or contact form) (
tormentor.me).
Who I’d actually contact
1) Sundeep Teki: He ticks most boxes for this brief. He lives and breathes “getting people into OpenAI/Anthropic/DeepMind” (
www.sundeepteki.org). His track record (100+ placements (
www.sundeepteki.org)) and explicit focus on those labs make him the obvious first choice. Even if he isn’t an alignment researcher, his coaching is laser-aligned with the
product (landing a job at OpenAI/Anthropic) Had cares about. In contrast, Tor’s profile is promising but more indirect to RLHF content. So Sundeep would be first priority.
2) Dr. Tor Bøe-Lillegraven: If I still had budget/time, I’d follow up with Tor. His coaching is broader, but his experience with AI safety/trust is unique. He could help with framing big-picture alignment thinking for interviews. He’s a long shot, but better than generic bootcamps.
Who I rejected and why
- Generic ML/AI coaches: Highly regarded firms and coaches like Exponent, Interview Kickstart, Lelding, Calibrant, etc., were immediately ruled out by scope. They are either broad, FAANG-focused, or AI-general “career coaches,” not specialized to frontier labs or RLHF. (For example, Exponent has ML practice but no visible RLHF content.)
- Leland/mentor-platform coaches (e.g. Parul S., Dan O. on Leland): These are individuals on marketplaces with solid ML backgrounds, but they serve broad audiences (career-changers, entry-level, general AI roles) and lack evidence of deep involvement with alignment or actual frontier labs. E.g. Parul Sahai’s profile shows general AI transition coaching to big tech, not specifically to Anthropic/OpenAI. I ruled them out for lacking the niche focus.
- “Ex-OpenAI” X: Many current/former OpenAI people (even if expert on RLHF) have no public coaching profile or are unreachable. Self-promoting coaches without verifiable lab ties (or just written a LinkedIn post) were also ignored due to lack of evidence of actual placement success.
- Executive/AIEthics coaches (e.g. John Packer): While interesting (Packer’s fintech/AI background is real), their coaching is aimed at business leadership and trust, not the nitty-gritty of technical interviews. He and others like him might talk about “strategic AI” in general, but they aren’t front-line lab interview coaches. So I passed.
Overall, aside from the two above, I found
no other real humans with both serious frontier-AI pedigree and a track record of advising job candidates. The field is dominated by either company-run programs or people without specific frontier focus, so most “near-misses” were rejected.
---
Sources
1. Sundeep Teki - Dr. Sundeep Teki | AI Career Coach for Frontier AI Labs | Oxford & UCL PhD
2. Sundeep Teki - Dr. Sundeep Teki | AI Career Coach for Frontier AI Labs | Oxford & UCL PhD
3. OpenAI, Anthropic & DeepMind Interview Guides | Research Engineer and Scientist Roles - Sundeep Teki
4. tormentor.me
5. tormentor.me
6. John Packer | AI Agent Engineer & Executive Coach
7. tormentor.me
8. tormentor.me