Yatimai/ai-watch
Autonomous AI news agent on LangGraph. Daily briefings in production since February 2026 via GitHub Actions
Machine Learning Engineer · AI Systems & Performance · Open Source vLLM contributor
Autonomous AI news agent on LangGraph. Daily briefings in production since February 2026 via GitHub Actions
A high-throughput and memory-efficient inference and serving engine for LLMs
Iterative self-improvement for Text-to-SQL: rejection-sampling fine-tuning, execution-verified, STaR-inspired loop
Unsupervised ML for ride-sharing demand prediction. K-Means / GMM / BIRCH clustering on 4.5M NYC pickups. GMM (k=25) achieves 98.1% success rate within 3km radius across 168 time windows.
End-to-end ML deployment for car rental pricing and delay analysis. Streamlit dashboard + FastAPI prediction API with MLflow experiment tracking. Gradient Boosting Regressor (R² 0.726). Deployed on Hugging Face Spaces.
Real-time fraud detection pipeline with Airflow orchestration (real-time + daily batch DAGs), Docker, PostgreSQL, and scikit-learn.
Production-grade data platform combining OLTP (PostgreSQL), OLAP (Snowflake), MongoDB, Kafka, Debezium CDC, Airflow, dbt, and Terraform.
Visual RAG system for financial document analysis using ColQwen2.5, Qdrant, Claude Opus/Sonnet
A safetensors extension to efficiently store sparse quantized tensors on disk
Transformers-compatible library for applying various compression algorithms to LLMs for optimized deployment with vLLM