Projects

VectraDB Rust, Tokio, gRPC (Tonic), REST (Axum), Sled, Protocol Buffers

Engineered a persistent vector database with HNSW, LSH, and Product Quantization for different similarity-search trade-offs.

CodeAnalyzer Pro Python, LangChain, ChromaDB, AST, Groq, Docker

Built a code-aware RAG system that parses repositories with ASTs and lets you inspect structure, relationships, and issues through natural language.

NetGPT: Network Traffic Optimization & Attack Detection Transformers, Networking, PyTorch

Developed a transformer-based AI system to optimize network traffic and detect malicious activity in real time, enhancing network performance and security.

Debian CLI Agent LLM, Qwen2.5, Fine-tuning, Shell, Python

Developed a lightweight command-line agent that interprets natural language into shell commands using a fine-tuned Qwen2.5-0.5B model, optimized for Debian environments.

DocuMind: RAG with Multi-Vector DB RAG, LLM, LangChain, FAISS, Groq

Built a retrieval-augmented generation pipeline with multi-vector document evaluation, context refinement, and hallucination reduction, optimized for technical document QA.

SecureVision Computer Vision, Anomaly Detection, PyTorch

Designed a secure computer vision pipeline for anomaly detection using spatiotemporal encoding and latent feature monitoring across frames.