Sanlabs is a boutique AI lab building custom AI solutions for workflows that are too complex, specific, or sensitive for standard off-the-shelf AI tools.
We work with messy real-world data across documents, spreadsheets, PDFs, scans, databases, email inboxes, and legacy software systems.
Our work includes document intelligence, AI agents, RAG systems, browser and computer-use workflows, software integrations, and full-stack applications with user interfaces, backend infrastructure, and deployment setup.
We use frontier models from OpenAI, Anthropic, and Google Gemini where they make sense. For highly sensitive data, we also build secure local, on-premise, and air-gapped deployments on private infrastructure or modern Apple Macs.
About
Sanlabs was founded by Tanay Soni, who has spent the last six years building AI infrastructure, LLM frameworks, agents, RAG systems, and enterprise AI tooling.
Before starting Sanlabs, he was an early engineer at deepset, where he co-authored Haystack, an open-source framework for building LLM applications. His work has covered pharmaceutical audit workflows, aviation maintenance data, financial regulatory document intelligence, and AI agents for logistics and freight forwarding.
Contact
Have a project we can help with? Send us a short note at tanay@sanlabs.ai. We'll take a look and tell you if it seems like a good fit.