# Judgment Labs

> Judgment Labs is a San Francisco applied-research startup building the continuous-improvement layer for AI agents. Its platform and open-source SDK, Judgeval, trace, evaluate, and monitor 'deep agents' - systems that reason, plan, use tools, and remember - turning messy production data into measurable, repeatable improvements. Founded in 2026 by three childhood friends out of Stanford, it raised $32M in combined seed and Series A funding led by Lightspeed Venture Partners.

- **Founded:** 2026
- **Headquarters:** San Francisco, United States
- **Founders:** Alex Shan (Co-Founder & CEO), Andrew Li (Co-Founder & Chief Scientist), Joseph Camyre (Co-Founder & CTO)
- **Team size:** ~24 employees
- **Products:** Judgment Platform, Judgeval (open source SDK), Agent Search, Agent Judge, Behavior Discovery
- **Notable:** Raised $32M in combined seed and Series A funding led by Lightspeed within roughly six months (announced May 12, 2026)., Built Judgeval, an open-source SDK for agent tracing and evaluation, published on PyPI and GitHub., Founders bring research pedigree from Stanford NLP (Stanford AI Lab), TogetherAI, and Datadog.

## Products & services

- **Judgment Platform** — The continuous-improvement stack for agents - trace long reasoning, tool use, and memory, then convert production data into better agents.
- **Judgeval (open source SDK)** — Open-source post-building layer for agents. Tracing and evaluation toolkit whose environment data and evals power agent post-training (RL, SFT) and monitoring.
- **Agent Search** — Query across agent trajectories at a behavioral level, beyond filters and input/output keyword search.
- **Agent Judge** — Cheaper, more accurate trajectory-level evaluators using harnesses and LLM-as-a-judge techniques.
- **Behavior Discovery** — Surfaces failure modes and usage patterns from unlabeled production trajectories.
- **AutoRubrics** — Automatically constructs and refines evaluation rubrics from verifiable signals.
- **Judgment MCP** — Integration with Claude, Codex, Cursor and other MCP clients for coding-agent workflows.

## Achievements

- Raised $32M in combined seed and Series A funding led by Lightspeed within roughly six months (announced May 12, 2026).
- Built Judgeval, an open-source SDK for agent tracing and evaluation, published on PyPI and GitHub.
- Founders bring research pedigree from Stanford NLP (Stanford AI Lab), TogetherAI, and Datadog.

## Latest updates

- **2026-05** — Closed $32M in combined seed and Series A funding led by Lightspeed Venture Partners to build the continuous-improvement layer for AI agents.
- **2026-05** — Announced plans to expand research and engineering teams in San Francisco and grow its forward-deployed engineering function.

## Links

- Website: https://judgmentlabs.ai
- LinkedIn: https://www.linkedin.com/company/judgmentlabs
- Twitter/X: https://twitter.com/judgmentlabs
- GitHub: https://github.com/JudgmentLabs

---

Profile page: https://yespress.io/judgment-labs
Published by YesPress — https://yespress.io
Last updated: 2026-06-14