Tomás Linde
Job seekerOpen to rolesML engineer moving into evals — looking for a first safety role
Lisbon, Portugal
About
Six years of production ML infra. Finished an intensive AI safety course this spring and now building open eval suites for tool-using agents.
Projects & outputs
AgentTrip
2026Sandboxed benchmark for detecting unsafe tool use in autonomous agents.
Trusted recommendations
Written by people whose organization affiliation has been verified on AegisNet. Candidates choose which recommendations to attach to an application.
Top decile cohort member; unusually good at turning vague threat models into concrete tests.
Priya Raghunathan · Northlight Fellowship Verified