Thinking Machines Lab
0 open Forward Deployed Engineer roles at Thinking Machines Lab, plus 1 closed. Curated with comp bands.
On the board since Jul 2026 · no roles open right now
Our take
Why Thinking Machines Lab
Lab-tier comp for the team customers actually meet — fine-tuning as the wedge, FDEs as the delivery mechanism, at Mira Murati's $12B-and-climbing lab.
- FDE type
- True forward-deployed
- Travel
- High
- Comp
- Top of market
- Stage
- Frontier lab
A good fit if you're…
- Engineers who want frontier-lab comp and customer-facing scope
- People drawn to fine-tuning, RL and model-customization as the core product
- Candidates who want a founding-era seat at a marquee lab (Mira Murati, ex-OpenAI)
- Those fine with the specific Tinker FDE listing being closed today — a benchmark and a watch, not an open apply
The lab and the leadership
Thinking Machines Lab, founded February 2025 by former OpenAI CTO Mira Murati, is a frontier lab staffed heavily from OpenAI (including OpenAI cofounder John Schulman). It raised a ~$2B seed in July 2025 at a $12B valuation led by a16z, with Nvidia, AMD, Cisco and Jane Street in. By September 2026 it was reportedly in talks at a ~$40B valuation.
Tinker: the product the FDE ships
Tinker is TML's fine-tuning API for open-weight models — you write the training loop, Tinker handles the distributed GPUs. It launched October 2025 and reached general availability (waitlist removed, vision input added) on December 12, 2025, supporting models like Qwen3-VL and Kimi K2. The FDE is the human layer taking that training stack into customer environments.
What "forward deployed" means here
True forward-deployed with fine-tuning as the wedge: FDEs scope training runs, build eval loops, and get custom models into production alongside customers. It's one of the few FDE seats where the product is the training stack itself, not inference — so the work skews toward RL, evals and model-customization depth rather than app plumbing.
Comp & leveling
The Tinker FDE seat posted $350K–$475K — unambiguously top-of-market and on par with other frontier labs, confirming TML pays lab-tier for the customer-facing role. Equity is pre-IPO in a company whose paper valuation has climbed from $12B toward a reported $40B — high-upside, illiquid, and priced accordingly.
The compute and stability picture
TML signed a multi-year Nvidia partnership in March 2026 for at least one gigawatt of compute, and employs roughly 100–200 people as of mid-2026. It's not without turbulence: cofounder Barret Zoph and Luke Metz left for OpenAI in January 2026 — a reminder that founding-era labs see leadership churn even as they scale.
Who thrives, and who bounces
Thrives: engineers who want frontier comp, fine-tuning and RL depth, and proximity to a marquee founding team. Bounces: candidates who need an open req today (the Tinker FDE listing is currently closed), want stability over founding-era chaos, or want remote — this is a high-travel, SF-based, top-of-market seat.
Questions candidates ask
Is the Thinking Machines FDE role open?
The specific Tinker FDE listing is currently closed. The page is retained for the comp benchmark ($350K–$475K) and likely re-openings as Tinker's customer base grows.
What would an FDE here actually do?
Embed with customers adopting Tinker — scope fine-tuning and training runs, build eval loops, and get custom open-weight models into production. The product is the training stack, not inference.
How does the comp compare?
The posted $350K–$475K is top-of-market, on par with other frontier labs. Equity is pre-IPO in a company valued at $12B (July 2025) and reportedly raising near $40B by late 2026.
Is it a stable bet?
It's extremely well-funded (~$2B seed, a gigawatt-scale Nvidia deal) but founding-era — cofounder Barret Zoph and Luke Metz returned to OpenAI in Jan 2026, so expect churn alongside the upside.
Analysis updated 2026-09-22
Open roles at Thinking Machines Lab
Compare Thinking Machines Lab
Others in the same tier
Free forever. Unsubscribe anytime.