I build RAG systems and intelligent applications — the retrieval pipelines, the LLM tooling, and the interfaces that wrap them. Currently shipping multimodal RAG and prompt-injection defenses.
B.S. in Computer Science from CUNY John Jay; starting an M.S. in CS · Machine Learning at Georgia Tech.
An ML-and-systems engineer is most useful where the model meets the interface. I work across that seam — research, engineering, and design — to ship things people actually use.
Production retrieval-augmented generation. Hybrid vector / graph / keyword search, re-ranking, eval harnesses, and prompt-injection defence at the agent layer.
Type-safe, performance-budgeted apps with Next.js and FastAPI. WebSocket streaming, Postgres + Redis, JWT/RBAC, Dockerised deploys, 500+ concurrent users.
Applied machine-learning research. Currently a multi-layer prompt-injection defence with a BIT classifier — 96% FP-rate reduction. Paper in preparation.
Six years of work, school, and side-projects, indexed by date. Filter by kind in the live build — work, education, project.
Graduate study specialising in machine learning — deep learning, NLP, and large-scale systems.
Architected a RAG-powered learning platform for 1,000+ students with ChromaDB + Cohere re-ranking. Shipped Flutter/FastAPI/Postgres microservices, real-time WebSocket chat with streaming GPT-4o, and i18n across 4 languages with RTL.
Undergraduate degree covering CS fundamentals, information security, and applied machine learning.
Production-grade multimodal RAG platform processing PDFs, images, audio and video in real time. Hybrid vector/graph/keyword search with 99.5% uptime, JWT + RBAC, automated CI/CD, 500+ concurrent users.
BIT classifier achieving 96% reduction in false-positive rate (40.2% → 1.5%) for prompt-injection detection. Multi-layer defence: input sanitisation, semantic analysis, agent-level isolation.
Reduced educator content-creation time by 60%. Next.js SSR frontend with JWT auth; NLP pipeline using NLTK/spaCy for context-aware question generation.
Tutored 100+ students in Python, data structures, algorithms, and calculus. Led weekly group sessions; raised average quiz scores by 15%.
Built NLP applications using BERT for quiz generation and text summarisation, accelerating ed-tech content creation by 50%. RESTful APIs with FastAPI/Django/Postgres; React + TypeScript dashboard for student analytics.