- Built a local Python MCP proxy that enforces launch-time filesystem policies for allowed paths, exact arguments, and call quotas before forwarding agent tool calls.
- Implemented serialized quota reservation, redacted append-only audit events, fail-closed error handling, and explicit tracking for completed and unknown outcomes.
- Blocked 12/12 unauthorized proposals while completing 10/12 attack tasks and 24/24 controls in simulation.
Projects
Selected security and backend projects showing implementation depth, automated verification, careful measurement, and clear limits on what each result supports.
- Built a 64-question, 24-group point-in-time CTI benchmark; constrained output raised evidence binding from 83.3% to 86.5% but lowered exact accuracy from 33.3% to 26.0%; a 144-response follow-up found no material schema benefit.
- Tracked CISA KEV/CSAF, NVD, CVE, MITRE ATT&CK, and vendor source provenance with SHA-256 hashes.
- Enforced point-in-time cutoff and source-authority rules with evidence spans, abstention, and human review.
- Automated typed-schema validation, split isolation, abstention, secret scanning, and privacy-safe aggregate checks.
- Secured 2 Flask services using 3-sample Pearson/MSE keystroke profiles, 20-character bcrypt fallback, 5-attempt/15-minute rate limiting, 60-second single-use codes, 1-hour JWTs, and atomic profile re-enrollment.
- Hardened login via CSRF, CORS allowlisting, JWT binding, callback validation, and server-to-server exchange.
- Contributed to an autonomous prediction-market platform integrating xAI Grok event detection, OpenRouter trading agents, and Solana smart contracts for on-chain market creation and trading at HackPrinceton.
- Built a 4-route Flask pipeline processing up to 40 X posts/min by weighting 7 signals and clustering 10+ posts.