Working across multiple projects spanning from the start of ML/AI pipelines to the finished product in finance, computer vision, LLM/RAG, and fintech — with a focus on systems that are clean, measurable, and built to run without daily oversight.
End-to-end MLOps system generating BUY/SELL signals for 45 IDX blue-chip stocks. XGBoost with 200+ engineered features, automated retraining, real-time Grafana monitoring, async FastAPI serving.
XGBoost FastAPI MLflow Grafana Docker CI/CD
Detect → then classify: a two-stage inference architecture that detects cars and people, then zooms in for license plates and faces. Outperforms single-pass detection on small/dense objects.
YOLOv8 PyTorch OpenCV Ultralytics
Retrieval-augmented generation chatbot that answers questions about my resume. Built with LangChain LCEL, FAISS, FastEmbed (ONNX), and Groq's Llama — evaluated with RAGAS metrics.
LangChain FAISS FastEmbed Groq RAGAS Streamlit
A passion project predicting the next opponent in Magic Chess Go Go using historical match patterns.
TypeScript Game Analytics
Open to remote full-time, contract, or freelance engagements in ML Engineering, Data Science, MLOps, or Full-Stack Development.