The problem
Teams deploying AI can't see why models behave as they do, so they can't debug, steer or trust them
What they do
Interpretability lab and tools to understand, monitor and edit the internals of AI models
Traction Customers incl. Microsoft, Mayo Clinic, Arc Institute, Rakuten; no revenue disclosed
Hiring trend
- Open roles
- 29
- Ops roles
- 4
- Headcount
- ~70
Team ~51 (Apr 2026) → ~69 (Oct 2026), est. +35%; 27 open roles (29 postings incl. 2 general-interest), nearly all SF
Hiring in
- San Francisco
- London
Latest news
Oct 2026
Partnered with Baseten on Project Beacon to detect unsafe agent behavior in open models ↗Funding
- Last round
- $150M
- Stage
- Series B
- Date
- Feb 2026
Led by B Capital
- Hiring+1.529 open roles on the live job board (10–29 roles = +1.5)
- Funding recency+1.5Series B · $150M, Feb 2026 (8 months ago): 6–12 months = +1, $100M+ round +0.5
- Traction+0.75Some growth: Customers incl. Microsoft, Mayo Clinic, Arc Institute, Rakuten; no revenue disclosed
- Red flags0None found in the last 24 months
- Total3.751.5 + 1.5 + 0.75 − 0 = 3.75 → Tempo (3.5–4.24)
Distance (Unproven) is a separate read on how long the lead can last. It doesn't change the score.
Distance
Unproven · edge: learns from use
Best-funded independent interpretability lab ($1.25B valuation) with marquee partners, but commercial adoption is still early and frontier labs build interpretability in-house.
Sources: Goodfire: Our Series B, $150M at $1.25B (Feb 2026) ↗ · Goodfire + Baseten partnership (Oct 2026) ↗
Last checked Oct 25, 2026