The idea came from something I kept watching unfold while leading the development of an autonomous delivery vehicle as founder and CEO of Udelv, Inc. Our mechanical engineers routinely spent tens of hours trying to pick the right material for vehicle parts — and often defaulted back to the handful they already knew. With over 150,000 engineering materials out there (composites, alloys, resins, ceramics, polymers), comparing them exhaustively is tricky. Defaulting to familiar materials quietly makes products heavier, pricier, more complicated to manufacture and less durable than they need to be. So I started building an AI tool to address the issue. Materiallab.ai lets engineers describe what they need in plain language and filter by numerical property constraints, combining semantic understanding with hard engineering data to surface materials they'd otherwise miss. What began as an internal tool for one team has become something I want every engineer to have. Whether you design EVs, medical devices, aerospace parts, or kitchen appliances, you shouldn't need to know all 150,000 materials to pick the right one. Would love your feedback — try it at materiallab.ai.

Material Lab

Agentic tool for optimal material selection (ATOMS)

Agentic tool for optimal material selection (ATOMS)