Summary
Alberto Mejía is a founder and research engineer specializing in AI safety with 11 years of engineering experience and a track record of turning research-grade ideas into production-ready systems. Based in San Francisco, he built Safety Harness to run adversarial, dangerous-capability evaluations across API and black-box web agents, producing reproducible evidence and mapping findings to compliance frameworks. His background spans ML infrastructure and fraud automation at Twitch—where he automated payout reporting and improved observability—and enterprise MLOps modernization at Citi, giving him deep operational expertise in deploying and validating models at scale. Alberto combines hands-on systems work (Go, Python, TypeScript, AWS) with rigorous test suites and deployment workflows, and he has a history of reducing manual toil through automation from internships at IBM to current startup efforts. He’s equally comfortable with low-level performance tooling and high-level safety metrics, an uncommon blend that helps bridge engineering, security, and compliance in real-world agent assessments.
10 years of coding experience
3 years of employment as a software developer
Regents Diploma with Advanced Designation, Electronic and Computer Circuitry, 3.7, Regents Diploma with Advanced Designation, Electronic and Computer Circuitry, 3.7 at Saunders Trades & Technical High School
Master's degree, Computer Science, 4.0, Master's degree, Computer Science, 4.0 at Georgia Institute of Technology
Bachelor of Science - BS, Computer Science, 3.55, Bachelor of Science - BS, Computer Science, 3.55 at Rensselaer Polytechnic Institute
Exploratory Major, 3.35, Exploratory Major, 3.35 at University of Connecticut
English, Spanish