I build end-to-end AI systems — from model training and optimization to inference APIs, mobile apps, and cloud deployment.
- B.A. in Computer Science & Mathematics, Bennington College (May 2025)
- Focused on applied ML, AI infrastructure, and full-stack engineering
- Interested in LLM systems, multimodal AI, speech/vision, and scalable AI products
- LLM applications: RAG, agents, tool calling, embeddings, and evals
- Real-time inference systems for speech and text
- AI-powered mobile and web apps
- Low-latency, cost-efficient deployment pipelines
- Cloud and GPU-backed production infrastructure
- Built a production voice AI system with Whisper ASR, diarization, and speaker verification
- Supported streaming and batch inference for web and mobile clients
- Built a validation rule engine that converts an Excel-based schema into dynamic UI, PDF, and XML output
- Implemented runtime validation, rule handling, and XML/PDF generation from a single source of truth
Languages: Python, TypeScript, JavaScript, Rust, C++
ML/AI: PyTorch, TensorFlow, Hugging Face, OpenCV
Data/LLM: PostgreSQL, pgvector, Redis, Pinecone, Weaviate
Frontend/Mobile: React, Next.js, React Native, Expo
Cloud/DevOps: Docker, Kubernetes, AWS, GCP, Terraform, GitHub Actions
2nd Place — MIT BTT-AI AJL 2025 (Kaggle)
Built dermatology models designed for stronger performance across Fitzpatrick skin types I–VI using group-aware sampling, reweighted losses, calibration, and ensembling.