In this episode of Unapologetically Technical, I interview Jacopo Tagliabue, the founder of Bauplan. Jacopo, with a rich background spanning cognitive science, basketball data analysis, and NLP innovation, shares his fascinating journey from academia to entrepreneurship. We trace his path from Italy to New York and beyond, delving into his early forays into data science long before it was a buzzword, including analyzing bike-sharing data and professional basketball statistics.
Jacopo provides a unique perspective on the evolution of data work, from rule-based NLP to the transformative power of LLMs. He breaks down the architecture of Bauplan, explaining its innovative approach to functions as compute units, data versioning with "Git for Data," and efficient Python packaging. We explore the challenges of managing data science teams and the critical importance of fast feedback loops and minimizing cognitive overhead in data workflows. Jacopo also shares invaluable insights into the realities of being a technical founder, including the importance of finding the right co-founders and navigating the complexities of building a startup.
Join us for an in-depth conversation with a seasoned technical leader who is revolutionizing how we work with data in the cloud, emphasizing simplicity, speed, and developer experience.
0:00:00 - Introduction
0:05:10 - Turkey and Breakfast
0:14:22 - Attending Yale
0:16:11 - Luck and Databases
0:18:14 - Time at Google
0:21:58 - How Not Fired
0:32:16 - PhD at Stanford
0:38:11 - University of Waterloo
0:43:39 - Joys of Teaching
0:52:09 - What is Kuzu?
1:10:42 - File Format
1:13:14 - Importance of Schema
1:34:48 - Query Planning
1:38:52 - Roadmap
1:41:39 - Choosing a License
1:44:19 - Origin of Name
1:45:33 - Making Difficult Decisions
1:49:02 - Getting Others to Code