project portfolio
AutoLabeler: ML-Driven GitHub Automation Suite
End-to-end automation tool for GitHub repositories that classifies and labels PRs/issues using fine-tuned NLP models. Integrated dashboard provides real-time activity visualization and CI/CD deployment analytics.
- Stack: TypeScript · Express · MongoDB · Docker · GitHub Actions
- Automation: Webhook listener triggers rule-based + ML pipelines
- Inference: Zero-shot classification via Hugging Face APIs
- Monitoring: Live EJS dashboard with query filters and metrics
Goodreads Companion Extension
Chrome extension enhancing Goodreads with NLP-based sentiment analysis, library lookup, and visualization features. Designed for privacy-first, client-side inference with a clean, reactive UX.
- Frontend: Vanilla JS + Plotly.js visualizations
- NLP: Lexicon-based sentiment classifier for reviews
- Integration: Open Library + WorldCat APIs via ZIP search
- Persistence: chrome.storage.sync for cached state
LangGraph Scholar
Distributed FastAPI backend that uses LangGraph to parse and summarize academic preprints. Implements a multi-agent architecture for concurrent retrieval, summarization, and evaluation with fault-tolerant state recovery.
- Stack: Python · FastAPI · Redis · LangGraph · T5-base
- Architecture: Asynchronous agent graph with retry handling
- Persistence: Redis session management for resumable jobs
- API: REST endpoints serving structured summaries
Conway Live
Real-time distributed implementation of Conway’s Game of Life using Go concurrency primitives and WebSockets. Optimized for low-latency broadcast rendering on HTML5 Canvas.
- Backend: Go (Gorilla WebSocket, Mutex-controlled state)
- Frontend: JS + HTML5 Canvas for dynamic visualization
- Concurrency: Event-driven model with goroutines
- Deployment: Dockerized with CI workflows
NoteNest
Cross-platform note-taking system with AI summarization and real-time synchronization. Designed for scalability, offline-first usage, and adaptive retrieval across devices.
- Frontend: React + Vite interface with modular state management
- Backend: Node.js + Express API with WebSocket sync
- AI Integration: Gemini-based note summarization
- Database: MongoDB + Mongoose models
WordNet Temporal Explorer
Java backend powering temporal-semantic visualizations of linguistic evolution using WordNet graphs and NGram data. Focused on high-performance indexing and caching for sub-second query responses.
- Stack: Java
- Data: WordNet lexical graphs + multi-decade corpora
- Frontend: JS-based visualization of usage trends
technical skills
programming: Python · Java · JavaScript · TypeScript · Go · SQL · C · C++ · R · HTML · CSS · LaTeX
frameworks: React · Next.js · Node.js · Express · TensorFlow · PyTorch · LangGraph · FastAPI · pytest
technologies: Git · GCP · AWS · Firebase · GraphQL · Redis · MongoDB · PostgreSQL · Docker · OpenCV
devops: GitHub Actions · CI/CD · Docker Compose · Vercel · Railway