Hi, I’m Mai, a staff engineer based in Spain. I’m an AI research engineer at XBOW, working on AI for offensive security.

I build robust AI systems, with security as the field I’ve decided to deep dive into: from identity security at Auth0 to LLM security in my PhD.

Here I write about AI engineering and cybersecurity: practical, no hype.

GitHub · LinkedIn

Experience

  • AI Research Engineer, XBOW · 2026–now
    AI for offensive security: safety, benchmarks and red teaming.
  • Zapier · 2025–2026
    LLM gateway serving 10M+ requests per day to 10+ internal teams, critical to Zapier’s day-to-day operations.
    Led a team through the migration, the technical requirements gathering and the comparison of LLM gateway providers.
  • Staff Software Engineer, AI teams, Phaidra · 2024–2025
    AI agents that control industrial facilities like data centres in real time.
    Reported directly to a director, led 10+ people across squads and initiatives, and worked closely with research.
  • Auth0 / Okta · 2021–2024
    Senior ML Engineer → Staff ML Engineer & Tech Lead
    Machine learning for identity security, from detection models to the MLOps platform that ships them.
    Joined as the 4th person on the team and ended up technically leading 3 teams, 12 people in total, reporting to a director.
  • Data Scientist, Glovo · 2020–2021
    End-to-end models for the courier app.

Education

  • PhD, Universidad Complutense de Madrid · 2023–now
    Started on fake news detection, restarting in 2026 on LLM security.
  • MSc in Mathematics and Computation, Universitat Oberta de Catalunya · 2019

Publications

Talks, Open Source & Community

…and more on GitHub.