benchmarks
How frontier models hold up on ultra-realistic professional work.
sherpalabs.aiCrawled Nov 16, 2026
How frontier models hold up on ultra-realistic professional work.
RL environments, benchmarks, and training data for finance and coding, built with experts.
Join Sherpa Labs’ confidential network of experts and get paid to design the problems frontier AI can’t solve yet.
megafund-bench-1: 150 tasks from one ultra-realistic megafund private equity deal world, graded against expert rubrics.
prod-bench-1, a private agentic coding benchmark of feature work in a large production monorepo: one-paragraph requests, graded by hidden tests.
Open roles at Sherpa Labs, in person in San Francisco.
Data Partnerships Lead at Sherpa Labs, in person in San Francisco. $180K–$260K + meaningful equity + performance bonus.
Growth Lead at Sherpa Labs, in person in San Francisco. $140K–$190K + meaningful equity + performance bonus.
Research Engineer, RL Environments at Sherpa Labs, in person in San Francisco. $300K–$400K + meaningful equity + performance bonus.
Strategic Project Lead, Coding at Sherpa Labs, in person in San Francisco. $175K–$250K + meaningful equity + performance bonus.
Strategic Project Lead, Finance at Sherpa Labs, in person in San Francisco. $175K–$250K + meaningful equity + performance bonus.
Sherpa Labs builds the RL environments, benchmarks, and training data frontier labs need for real professional work.
Sherpa Labs builds ultra-realistic RL environments, benchmarks, and training data for frontier AI labs.
How Sherpa Labs handles personal information on sherpalabs.ai.
StoryHouse Review shifts focus from covering breaking tech news to investing in early-stage startups. The publication now writes its first checks as a venture fund, targeting seed-stage companies. The move marks a transition from editorial content to active participation in the startup ecosystem.