Senior GenAI Engineer — I build production LLM systems that survive contact with the enterprise.
4 years shipping agentic architectures, multi-hop RAG pipelines, and evaluation frameworks at scale. Currently at Deloitte (finance AI), previously RevGen Partners. Actively building in public.
| Project | What it does |
|---|---|
| ReasonForge | Modular math toolkit for LLMs — 111 tools, 7 MCP servers, exact computation via SymPy/NumPy. Listed on awesome-mcp-servers. |
| UpTier | MCP-powered task manager that lets Claude AI prioritize your work. Desktop app with Eisenhower matrix, quick-win scoring, and goal tracking. |
| DeckSnag | CLIP-powered slide capture — auto-detects slide transitions in video and exports to PowerPoint, PDF, or images. |
| TransitGuard | Predictive safety intelligence for Chicago CTA. XGBoost + DBSCAN on 50K+ crime incidents with RAG chatbot interface. Northwestern MSDS capstone. |
- LLM-as-a-Judge evaluation framework - reduced functional testing time by 90% across production AI outputs
- Finance agents on Google Gemini Agentspace - agentic workflows with human-in-the-loop oversight for high-stakes enterprise contexts
- Multi-hop RAG architecture - multi-modal document processing on GCP (Vertex AI, Document AI), optimized for latency and cost at scale
- ML lead routing engine - ranking, ensemble, classification, and forecasting models deployed across enterprise client engagements
LLM / Agents: Python · LangChain · LlamaIndex · OpenAI APIs · Vertex AI · AWS Bedrock · MCP
Infra / Data: GCP · Azure · PySpark · SQL · Airflow
ML: Scikit-learn · XGBoost · TensorFlow · Hugging Face
Tooling: Docker · FastAPI · Streamlit · TypeScript
🎓 MS Data Science (AI Specialization) - Northwestern University
📜 Azure Data Scientist · Databricks LLM · IBM Data Science · Professional Scrum Master
📍 Denver, CO