Data from the Physical World

about this room

Two engineers on what happens when machine learning meets greenhouses and building sites: sparse, hand-recorded data, plants that ignore your model's assumptions, and a robot fleet you have to debug from 100 km away.

PyData Amsterdam with Monumental and Source.ag. At 18:30 Sebastiaan Vermeulen, who leads harvest forecasting at Source.ag, explains why a season gives you maybe 40 data points per crop and how mechanistic biology models are blended with learned ones. At 19:00 Josefine Quack, forward-deployed robotics engineer at Monumental, on continuously deploying to more than 100 brick-laying robots working live sites. Around 30 minutes each with questions, then dinner and drinks until 21:00. Walk-in from 18:00.

Free. PyData Amsterdam has moved registration to Luma; sign up there rather than on Meetup.