What I'm Working On

Side projects and things I build outside of work.

Interview Command Center

An AI-powered interview coaching system that transcribes recordings, analyzes performance across rubric dimensions, and generates actionable prep briefs. Uses a custom NeMo Parakeet pipeline for GPU-accelerated transcription.

PythonGCPNeMoLLM
More
Process Insights — AI-generated assessment with advance/pass/fail recommendation
Process Insights — AI-generated assessment with advance/pass/fail recommendation
Question Bank — 243 questions tracked with filtering by type and source
Question Bank — 243 questions tracked with filtering by type and source
Top Strengths and Improvements — side-by-side analysis with talk time ratio
Top Strengths and Improvements — side-by-side analysis with talk time ratio
Coaching Tips — actionable recommendations for interview technique
Coaching Tips — actionable recommendations for interview technique
Reviews — diarized transcript playback with timestamped speaker labels
Reviews — diarized transcript playback with timestamped speaker labels

Automated Prediction Market Arbitrage Engine

A real-time arbitrage engine that identifies pricing discrepancies between Polymarket and Kalshi prediction markets. Built in Rust for low-latency execution with gRPC streaming and automated position management.

RustgRPCPrediction Markets
More

Screenshots and examples coming soon.

Writing Style Kit

A model-agnostic system for capturing and reproducing your personal writing voice across any LLM. Ingests a curated corpus tagged by register, distills a structured style guide, and pairs it with embedding-based few-shot retrieval at generation time. An MCP server exposes your voice as tools any compatible client can call. Building toward a custom fine-tuned model with PyTorch.

PythonPyTorchMCPRAG
More

Screenshots and examples coming soon.