Data Science · Machine Learning · Data Engineering · AI Engineering
I build end-to-end systems where data, machine learning, and modern AI meet.
From pipelines and explainable models to LLM agents, retrieval, evaluation, and production workflows.
- AI systems: agents, RAG, embeddings, tool calling, MCP, Text-to-SQL and context-aware workflows
- Reliable LLM workflows: evaluation, guardrails, verification, tracing, memory and model routing
- Data + ML systems: pipelines, analytics, applied machine learning and model explainability
- Production engineering: Azure, Docker, Terraform, CI/CD, MLOps and observable services
I like understanding what happens under the hood, building the important parts from first principles when it helps, and then making the system reliable enough to measure and use.
AI & Agentic Systems
LLMs · AI Agents · RAG · MCP · Tool Calling · Text-to-SQL · Embeddings · Evaluation · Guardrails
Data & Machine Learning
Python · SQL · PySpark · Airflow · Machine Learning · XGBoost · PyTorch · SHAP · Tableau
Platform & Engineering
Azure · Docker · Terraform · GitHub Actions · FastAPI · Git · Linux
A text-to-SQL agent built from the loop up instead of hiding the behavior behind an agent framework.
It explores tool calling, retrieval, plan → verify → repair, SQL guardrails, memory, tracing, evaluation, embeddings, model routing and MCP against a real analytical warehouse.
LLM Agents Text-to-SQL Retrieval Evaluation Tracing MCP Azure
|
PySpark, Airflow, Azure and end-to-end ML pipeline engineering. |
Forecasting with XGBoost, SHAP explainability and Tableau. |
NLP with PyTorch, CNNs and LSTMs. |
late nights · curious systems · build from first principles · ship what works