AI evals nonprofit Sophron Research hiring founding scientist at $160K–$300K
eli_lifland · x · 2026-09-24
Sophron Research, a new independent nonprofit building evaluations for frontier AI models, is hiring a Founding Research Scientist (NYC or remote, $160,000–$300,000).
- Mission: empower human judgment in a world of advanced AI; first eval focused on model sycophancy (Pander Score)
- Planned projects: multi-turn Pander Score evals, oversight evals for accuracy and legibility of task reports, pro-developer bias in model answers/reasoning, and how models represent uncertainty in behavior vs. statements vs. internals
- Startup-style role with substantial research autonomy, including stakeholder communication with labs and regulators
- Researcher Eli Lifland endorsed the role, stressing that evaluating how AIs shape people's beliefs is very important, with future evals possibly probing whether models favor their creators
More from Companies & People
- Meta unveils $1,299 VR glasses: 5K micro-OLED, no head strap, shipping Spring 2027 — beffjezos · 2026-09-24
- Meta's Muse Connector Platform draws 2,000+ developer submissions days after launch — Rasmic · 2026-09-24
- Even Grok ships models faster than Google — the 80,000-engineer problem — bindureddy · 2026-09-24
- Durable's solo founder: AI site builder hit 1M users in 3 months, now 3.5M SMB signups — julianweisser · 2026-09-24
- Education's Twin Missions: Backing Proven Talent and Unpolished Diamonds — anshulkundaje · 2026-09-24
- Three years ago Claude was a punchline: Garry Tan amplifies lesson on iterating to PMF — garrytan · 2026-09-24