I turn complex machine learning, biomedical signals, retrieval pipelines, and raw business data into reliable models, grounded AI workflows, decision-ready dashboards, and deployable services that help teams evaluate evidence, understand performance, and act on insights with confidence.
| Area | Technologies |
|---|---|
| Programming & ML | Python, SQL, R, scikit-learn, XGBoost, LightGBM, pandas, NumPy, SciPy |
| Data analysis & BI | Power BI, Tableau, Microsoft Excel, Streamlit, Matplotlib, Seaborn |
| Data engineering | Apache Spark, PySpark, PostgreSQL, MongoDB, MySQL |
| Deep learning & biomedical AI | PyTorch, TensorFlow, Keras, VGG16, Swin Transformer, MNE-Python, YASA |
| Generative AI & RAG | LangChain, LangGraph, LangSmith, RAGAS, Sentence-Transformers, Pydantic, OpenAI SDK, Ollama |
| Vector search & retrieval | FAISS, Annoy, ChromaDB, PostgreSQL/pgvector, OpenSearch |
| Backend & APIs | FastAPI, SQLAlchemy, Redis, Celery, httpx |
| MLOps & cloud | Docker, Docker Compose, DVC, MLflow, GitHub Actions, AWS EC2, AWS ECR, Amazon S3, Amazon CloudWatch, Google Cloud Platform |
Across these projects, I focus on:
- validating data quality and defining clear problem requirements before modeling;
- selecting evaluation methods and metrics that reflect the objective and intended use case;
- building reproducible workflows for data preparation, experimentation, analysis, and model development;
- packaging analytical and machine-learning solutions so they can be tested, maintained, and reused;
- documenting assumptions, limitations, results, and technical trade-offs alongside successful outcomes.