How does bauplan work?

Think of an Operating System (OS). It lets you run multiple applications at once by managing resources like I/O, memory, and CPU in the background. The OS allocates, frees, and prioritizes resources, ensuring smooth communication between hardware and software.

bauplan works in a similar way. Our proprietary runtime allows us to run Python and SQL jobs as ephemeral functions that break down data workloads into smaller functions and decides how to allocate resources to compute them using the same resource pool.

This is very different from traditional big-data engines, like Spark, that encourage monolithic processes and require separate clusters for different workloads.

We spent a lot of time optimizing our runtime for subsecond container spin-ups, efficient object storage access, smart caching and in-memory processing. As a result, we are up to 100x more cost-efficient and require zero management from data platform teams.

Team

We are a small team of engineers out of New York and San Francisco. After years working in AI, we became passionate about making it simple for every developer to work with data infrastructure. We value strong technical skills, frugality, deep customer empathy and adventurous offsites.

Ciro, Jacopo and Mattia are serial entrepreneurs who went from early stage to acquisition to IPO in less than 5 years, and our founding team comes from Docker, Amazon and top academic institutions.

bauplaners have a few Ph.D.s, 40+ research papers (including Nature, ICML, KDD, VLDB, SIGMOD), thousands of citations and several popular open source contributions, with millions of downloads and >10k GitHub stars.

Bauplan Team

Oct 2024