Timeline

Graded real-world GitHub PRs to curate RLHF data, directly training the automated code review architecture launched in Claude Opus 4.6.

Architected secure, containerized Docker sandboxes for execution-based AI evaluation. Engineered isolated Linux environments to benchmark the autonomous CLI and coding capabilities of frontier models.

Contract work across model evaluation, safety, and coding-agent quality for multiple AI teams. Client details are included where permitted.

Audited API execution logs and enforced factual grounding to mitigate hallucinations in Gemini 2.0's agentic tool-use capabilities.

Leveraged internal WFE access to conduct adversarial safety evaluations on Meta AI.

Undergraduate studies in computer science and engineering.

Completed higher secondary education at Notre Dame College.