Catch up on the latest talks from Munich Datageeks — fresh insights, cutting-edge tech, and real-world data stories from top minds in the field. Stay curious and stay inspired!
Learn more about the hottest topic's in data and AI.
Stop building agents like interns. Nicholas (Superglue) argues that fewer, general-purpose tools — terminals, CLIs, file systems — dramatically outperform large curated toolsets, backed by examples from AWS, Vercel, and Anthropic.
Storm Reply and AWS present a serverless AWS pipeline for extracting clinical entities from German healthcare documents, combining Textract, Claude Sonnet, Medical Comprehend, and Claude Opus — cutting doctor pre-consultation review to under two minutes.
Patricia (VMO) presents three automated data governance solutions for Metabase: HR-driven permission sync via Personio & Dagster, dbt YAML-based documentation sync, and schema-change-resilient query updates using memoized API calls — all built with Python.
BSH applies LLMs to systems engineering — automating requirements syntax checking, quality validation, and rewriting — while building a centralized AI platform to standardize RAG pipelines, observability, and governance across 400+ enterprise use cases.
ML project failures—from misaligned goals to ignored baselines—stem from one root cause: lack of design documentation. Six real-world cases demonstrate how two weeks of upfront doc writing prevents months of wasted effort and embarrassing failures.
