
Using publicly available information we constructed an analysis to help you get a feel for this speaker before deciding to attend their session.
Marina Petzel spent ~8 years in data science and shipped customer-facing ML at Autodesk — most notably AutoCAD Smart Blocks, a geometry-recommendation feature taken from prototype to production in a 40+ year-old product — and is now a GenAI advocate at Datadog. Her recurring practitioner thesis: roughly 70% of real ML work is data collection, cleaning, and pipelines rather than model training ("garbage in, garbage out"), and AI is most useful when embedded into workflows rather than bolted on.