1
Ask HN: Where's the actual pain in early-stage medical AI startups?
I'm a 46 year old ML engineer (PhD + MD, NVIDIA edge stack) exploring solo consulting in medical AI. Did a deep dive on the market. Short version: regulatory expertise is valued, pure ML is oversupplied, consulting window closes at Series B.
What I couldn't find: strong signal on what early-stage teams actually struggle with.
Background: years on Jetson/TensorRT, classical CV in health and Earth observation, paper-to-production work at startups. I'm targeting EU companies (US market is largely closed to remote international consultants), but asking here because US is ahead. Your current pain is likely EU future pain.
What's eating your runway right now? Not "we need data." The specific technical or non technical friction blocking your next milestone.
Curious about CV-heavy use cases, edge deployment on constrained devices, anything touching imaging or surgical visualization. Wouldn’t go to LLM (eg RAG on patient data) w/o very strong signal that this is the only niche left. Also, not much experience on graph networks (drug discovery or protein folding).