Company
Latent catches the answers a model makes up.
We build the check that reads a language model while it writes and sends the answers it cannot vouch for to your reviewers.
Built by people from
Research published at
What Latent is
Latent catches the answers your model makes up and acts on them before a customer sees them. It reads a model's internal state: directly as it writes, when Latent runs beside an open-weight model on vLLM, and through Latent's own reader model for every other model, closed models behind an API included. When it flags an answer, Latent delivers it and logs it, holds it with your text, or hands it to a stronger model, and a flagged stream is cut at the first unsupported sentence. It runs hosted by Latent or inside your own environment.
Founder
Models leave a trace when they make something up, and I built Latent so the teams who stand behind every answer can catch it before their customers do.
Published LLM research at ACL 2023, EMNLP 2025 and NeurIPS 2025. Previously did research at Stanford and Princeton across reasoning, evaluation and interpretability, and built ML at Robinhood and Amazon. Dual degree in computer science and finance from Penn's M&T program (Wharton).
How to reach us
- Book a call
- latent.cal.com/vedant
- Security review
- The security detail
- Vulnerability reports
- Disclosure policy