DuckDB 2.0 + Data Infrastructure for Science!

AGENDA

• 6-6:30pm - Food and Networking
• 6:30pm - Introductions
• 6:45pm - Talk #1: DuckDB 2.0
• 7:20pm - Talk #2: Data Infrastructure + Science
• 8:00pm - End

Talk #1 Abstract: DuckDB 2.0 is now available! This talk will cover some of the new features and performance improvements that make DuckDB even more awesome.

Talk #1 Speaker Bio: Will Angel - ML Engineer @ Procore

Talk #2 Abstract: Search Is Still Unsolved: Open-Source Data Infrastructure for Undiscovered Public Knowledge on Standard Consumer Hardware

Our collective knowledge infrastructure - textbooks, training resources, literature synthesis - is quietly accruing a structural liability: compounded confirmation bias, stacked citation-by-citation into the foundations of formal knowledge, means breakthroughs take decades to reach the classrooms and clinical workflows where they matter. The deepest friction is that before a field can agree on "why," it must first agree on "what" has been observed - a step routinely skipped or fragmented across silos that never cross-pollinate. This talk argues that fixing this is a data engineering problem, not a philosophy problem. By composing tools already at our disposal - LLMs, classical NLP, public data repositories, and automation frameworks - we can model knowledge itself rather than imitate individual experts: algorithmically mapping what has been said and measured across a problem space, separating observation from interpretation, and automating the writing of "living textbooks." The deeper aspiration: systematically surfacing hypothesis combinations that no siloed researcher would have the cross-disciplinary vantage point to ask. Demoed with ME/CFS research, a use case showing what siloed, fragmented knowledge infrastructure costs in practice - and what a data engineer's toolkit can do about it.

Talk #2 Speaker Bio: Sam Keating is a Principle Investigator at DMV Petri Dish, Systems Biologist, and published Citizen Scientist