Projects

An open-source scanner that attacks your RAG pipeline for prompt injection and document leakage, and fails CI the moment it gets more exploitable.

On PyPI and the GitHub Marketplace

An independent reimplementation of MINJA, a memory-poisoning attack on LLM agents that needs nothing but ordinary queries, plus the confidence intervals the paper left out.

Reproduction of arXiv:2503.03704

A Kubernetes remediation tool that holds no cluster credentials. Its only write is a git commit, so every fix it makes can be undone with git revert.

Technical report and live demo

Watches Android security, CVE and policy feeds, works out what actually changed, and turns the significant changes into cited risk tickets.

Team capstone with TransUnion

A GPU-accelerated document-intelligence platform for CCC Intelligent Solutions that grounds every answer in the exact source paragraph.

Client project for CCC Intelligent Solutions

A controlled test of whether chain-of-thought supervision helps a small vision-language model reason about scenes, using BLIP-2 fine-tuned with LoRA.

CS533 Deep Learning for NLP, team of 2

A team course project that served an SVD++ recommender through FastAPI with model routing, per-request provenance, data checks and CI.

Responsible AI Engineering, team of 6

How well small local LLMs check FEVER claims with and without evidence, and how easily hand-written counterfactuals fool them.

CS421, team of 4

A fuzzy-logic language embedded in Scala 3, with fuzzy sets, logic gates, scoped variables, classes and partial evaluation.

CS476 Programming Language Design

A chat app with MTProto 2.0-style message encryption on the server and an end-to-end encrypted secret chat in the browser.

CS588 Security and Privacy in Networked and Distributed Systems

My build-scan-deploy pipeline and AWS setup for running a community three-tier app on EKS, with Terraform, Jenkins, SonarQube, Trivy, ECR and an ALB ingress.

Built on a community reference app