ML Engineer. Builder of agentic AI systems and production LLM pipelines.
MS Data Science @ Northeastern University (Khoury College)
I care about shipping ML systems that work for real users. Most of my work lives at the intersection of LLMs, agentic workflows, and production infrastructure. I've built everything from fine-tuned language agents to adversarial robustness tools to full-stack hiring pipelines powered by LLM orchestration.
Currently a TA for AI-Assisted Software Engineering at Northeastern, co-founder of an AI productivity startup (Twinly), and always building something new.
CodeRadar MCP Server - A Python MCP library for local codebase intelligence. Repo explanation, PR diff analysis, dependency CVE auditing via OSV.dev, static analysis with ruff/pyflakes, and self-healing code suggestions. Designed for PyPI publication.
- BERT-Based Treatment Recommendation System - First-place international award, published in IGI Global (2024). Built a BERT-powered system for personalized treatment recommendations.
- Adversarial ML Resilience Against CNN Data Poisoning - Peer-reviewed research on defending convolutional neural networks against data poisoning attacks.