Resume
Education
BSc, AI Engineering — Universidad Francisco de Vitoria, Madrid. 2024 – Present. Second-year undergrad in Sistemas de Inteligencia Artificial.
High school — Trinity College, San Sebastián de los Reyes.
Experience
Production LLM agents (2026 – Present). Shipped commercial Salesforce integrations, MCP servers with security-review loops, Android field assistants, multilingual institutional tooling. Small independent client base + university projects.
Research pivot (2026 – Present). Moving from shipping production systems toward evaluation and alignment research. So far that has meant contributing to the open-source eval frameworks that safety organizations actually run, listed below.
Open source
3 PRs merged in eval / agent-tooling frameworks:
- princeton-pli/hal-harness #182 — weave/gql dependency alignment.
- METR/hawk #627 — whitespace-only credential handling in the passthrough middleware.
- affaan-m/ECC #920 — token-budget-advisor skill for Claude Code.
See GitHub for the current list.
Currently learning
Eval methodology and agent harnesses, mostly by working inside them (HAL, hawk, Inspect).