Building intelligent systems at the intersection of AI, high-performance computing, and full-stack engineering. MS CS @ Rutgers, graduating Fall 2026.
About
I'm a CS graduate student at Rutgers focused on the full spectrum of intelligent systems — from training spiking neural networks on real industrial sensor data, to running LLMs natively in the browser via WebGPU with zero API cost. I care about bridging the gap between research and production.
On co-op at Greenlit, I've shipped AI-powered genre market analysis tools from scraping to a FastAPI service used in production decisions. Outside of work, I'm usually benchmarking something, building something, or both.
Currently exploring Rust, WebAssembly, and Reinforcement Learning.
Career
AI-powered entertainment investment platform
State University of New Jersey
Student-run product management fellowship at Rutgers
Work
Full-stack auction platform with role-based access for admins, customer representatives, and buyers. Features live bidding with reserve-price enforcement, preference-based email alerts, image uploads, and an admin analytics dashboard.
↳ 3-role system · 20+ pages · preference-matched email notifications on new auctions
Browser-native LLM benchmarking tool that runs small open-source models entirely via WebGPU — zero server, zero API cost. Measures TTFT, tokens/sec, thermal throttling, and MMLU accuracy, then compares against cloud APIs (Gemini, Claude) on cost and quality.
↳ Llama 3.2 1B: 73–172ms TTFT · 48–52 TPS · $0 vs. measurable cloud API cost per 20 queries
Brain-inspired computing research project comparing CNN and Spiking Neural Network architectures for anomaly detection on industrial bearing vibration data (CWRU) and cardiac ECG signals. Implements CNN→SNN weight transfer and surrogate gradient training across 3 architectures × 2 datasets.
↳ 6 notebooks: CNN→SNN weight transfer + surrogate gradient SNNs on CWRU bearing & MIT-BIH ECG
Explainable churn prediction pipeline trained on three real-world datasets — Telco (~7K rows), SaaS (~963 rows), and BankChurners (~10K rows). Compares Logistic Regression, Random Forest, and XGBoost with dataset-specific hyperparameter tuning, then explains every prediction via SHAP beeswarm and bar plots.
↳ AUC ≥ 0.85 on Telco & BankChurners · SHAP beeswarm + bar plots per dataset per model
AI research project implementing robot self-localization under uncertainty across 4 phases: optimal BFS over belief-set state space → greedy strategy → CNN regressor trained to predict localization cost → 3-policy comparison. The agent navigates a maze without knowing its position, narrowing a belief set to a single cell.
↳ CNN regressor (π₁) outperforms greedy baseline (π₀) · dual-policy π₂ evaluated across belief sizes 1–20
Native Android photo management app built in Java with 11 Activities covering the complete feature set: create/rename/delete albums, add and tag photos from device storage, search across all albums by Location and Person tags with autocomplete, slideshow view, and on-device serialized persistence.
↳ 11 Activities · AND/OR tag search with autocomplete · zero third-party dependencies
Fully playable two-player console chess engine in Java implementing the complete rule set: all 6 piece types with legal move validation, check and checkmate detection, en passant, castling with first-move tracking, pawn promotion to any piece, draw offer, and resign.
↳ Full rule set: en passant · castling · checkmate detection · pawn promotion
Interactive Experiments
Click play on any card to run the demo directly in the browser — no install, no API keys.
Real measured latency and throughput data from running Llama 3.2, Qwen 2.5, and SmolLM2 in the browser via WebGPU across three hardware profiles.
Step through pre-scripted move sequences on a static CSS/SVG board demonstrating en passant, castling, and checkmate from the Java chess engine.
CSS keyframe animation of neurons firing in sequence — visualizing how spiking neural networks propagate activation during fault detection inference.
Static frontend snapshot of the auction platform — a fake live auction with a countdown timer, animated bid feed, and reserve price indicator.
Stack
A toolkit spanning AI research, systems engineering, and full-stack development.
Currently exploring Rust, WebAssembly, and Reinforcement Learning — feel free to ask about those too.
Portfolio Search
Search by role, technology, or project type. Answers come only from local portfolio data.
This assistant searches only local portfolio content — no external AI API, no backend, no API keys.
Get in touch
Open to full-time opportunities, internships, and interesting engineering conversations.