Remote Jobs at Collinear AI

Collinear AI

11 - 50 employees

Founded 2024

🤖 Artificial Intelligence

🏢 Enterprise

☁️ SaaS

Artificial Intelligence • Enterprise • SaaS

Collinear is a company that provides a platform for evaluating and improving frontier AI agents through realistic, multi-turn simulation labs. Collinear packages off-the-shelf tasks and rigorous evaluation (including deterministic verifiers and rubric-scored outcomes) so teams can repeatedly exercise agents in self-contained simulated worlds that include users, tools, persistent state, and realistic enterprise scenario data. Their product offers plug-and-play agent integration, stateful sandboxes and cloned enterprise tools, non-scripted simulated users, and a broad catalog of tasks (e. g. , Cyber SWE, Computer Use, Knowledge Work, Behavior). Collinear publishes focused benchmarks and labs such as CWE-bench for cybersecurity (4,000+ tasks, 215 distinct CWEs, 120 held-out eval tasks, coverage across C/C++, Go, Java, TS/JS, Python, Rust; reports like 71% pass rate for frontier models and claims of 40%+ measured performance lift, 90% functional fidelity with real products/tools, and 500B+ tokens of agent training data generated). The platform is aimed at helping enterprise teams maximize measurable capability gains from each model iteration by creating realistic, repeatable training and evaluation environments.