When FaaS met data: Challenges and opportunities
Abstract:
Data practitioners today spend a significant fraction of their time dealing with infrastructure, glue code and implementation details, making data pipelines slow to build and costly to maintain. While pipelines are purely functional, current FaaS platforms are ill-suited for data workloads: in particular, moving data between functions should not be a developer concern, but a platform optimization. In this talk, we describe the technical choices we made when starting Bauplan - a new serverless data platform -, and discuss preliminary results from optimizing data sharing with Arrow.
Shortbio:
Jacopo Tagliabue is the CTO and co-founder of Bauplan. Educated in several acronyms across the globe (UNISR, SFI, MIT), Jacopo was co-founder of Tooso, an AI startup acquired by Coveo. He led Coveo’s AI roadmap from scale-up to IPO, and built out Coveo Labs, an R&D practice rooted in open source and open science. While building his new company, Bauplan, he moonlights as Machine Learning Systems Professor at NYU, which is mostly notable because it is the only job he ever had that his parents understand.