I’m a systems & data infrastructure founder based in Geneva, Switzerland 🇨🇭 (originally from Porto, Portugal 🇵🇹).

I lead RAW Labs, where we’re building AI‑native data platforms.

• MXCP - An open-source, declarative framework for building and running AI tools & APIs (powered by the Model Context Protocol)
• AI engineering: agent SDKs (e.g. PydanticAI), tool architectures, execution runtimes
• Data systems: DuckDB, dbt, PostgreSQL (incl. FDWs), streaming connectors
• Systems & platforms: Python runtimes, sandboxing, secure code execution (e.g. MicroVMs, LSM), and everything needed to run AI workloads safely at scale
• And of course, plenty of Python, and whatever else helps businesses make sense of their data

I (sometimes) publish notes, prototypes, and reflections at notes.spacebase.org.

My background is in distributed systems and databases.

Nowdays, have a growing focus on AI systems, particularly at the intersection of data access, analytics, and data safety.

Before founding RAW Labs, I published on topics like scientific data processing and large-scale distributed data management. Check out my Google Scholar profile for more.

• CEO & Co-Founder – RAW Labs Mar 2015 – Present | Lausanne, Switzerland

Bringing the next-gen data infrastructure to market: a SQL engine built for declarative AI, modern data APIs, and real-time analytics — all over raw, remote, and nested data.

• Researcher in Database Systems – EPFL Oct 2010 – Feb 2015 | Lausanne, Switzerland

Co-creator of RAW, a novel database architecture for high-performance querying of raw data. Researcher in the Data-Intensive Applications and Systems Laboratory.

• Software Engineer – CERN May 2003 – Mar 2009 | Geneva, Switzerland

Software Engineer for the ATLAS Experiment at the Large Hadron Collider. Led the architecture and development of the ATLAS Distributed Data Management system, which handled over 15 petabytes of scientific data.

• PhD, Computer Science (2004 – 2009), University of Southampton Researched large-scale distributed systems and managed petabyte-scale scientific datasets in grid environments.

• Engineer, Software Engineering (1998 – 2003), Universidade do Porto

• Scala, Python, C, Rust, TypeScript, Bash
• PostgreSQL, DuckDB, gRPC, Protobuf
• Kubernetes, GitHub Actions
• macOS (with a flavor for CLI minimalist setups)
• .. and lots of Cursor and Replit, lately.

• Twitter/X: @miguelbranco80 (don't use much...)
• GitHub: You're here already
• Blog: notes.spacebase.org
• Website: spacebase.org

“Simple is better than clever. Declarative is better than obscure.”

"Systems, not hacks."