I build data pipelines, machine learning systems, and decision intelligence tools for real-world applications.
- 🔭 Currently working on Machine Learning & Data Engineering systems
- 🌱 Exploring MLOps, real-time prediction systems, and AI agents
- 🤝 Open to collaboration on ML systems & backend data infrastructure
- 📊 Interested in predictive analytics and optimization problems
-
📈 Uplift Modeling MLOps Pipeline Production-style ML system for causal uplift modeling using the X5 RetailHero dataset, featuring FastAPI deployment, Docker containerization, Kubernetes orchestration, CI/CD automation, and real-time monitoring with Prometheus and Grafana
-
🚉 Real-Time Transit Delay Prediction System End-to-end ML system with real-time ingestion, prediction API, and monitoring
-
🧪 DataForge Bench — ML Data Format Benchmarking Platform Interactive system benchmarking JSON, CSV, Parquet, Protobuf, and Avro across full ML pipelines, evaluating serialization, storage, streaming, feature engineering, and training performance with optional Python backend validation
-
⚡ EV Charging Demand Modeling (Research) Graph-based zoning + demand modeling using arterial networks
-
🤖 AI Outreach Agent Automated lead generation, enrichment, and communication
Languages: Python, SQL ML/AI: Scikit-learn, LightGBM, Feature Engineering Data Engineering: Airflow, Kafka, ETL Pipelines Systems: FastAPI, Docker, APIs Other: NetworkX, Optimization, Analytics
- 🌐 Portfolio: https://ehsansobhani.vercel.app
- 💼 LinkedIn: https://www.linkedin.com/in/ehsan-sobhani-858b453b/
- 🐙 GitHub: https://github.com/ehsansobhani
Building production-grade ML systems, not just models.