Website: https://ben-geo.github.io/
LinkedIn: https://www.linkedin.com/in/bengeoabraham/
GitHub: https://github.com/Ben-geo
Email: [email protected]
I’m an Applied Machine Learning Engineer focused on building scalable, high-performance ML systems and data platforms.
My work spans reinforcement learning, LLM-based tools, and large-scale data engineering, with a focus on research-driven engineering, system optimization, and production reliability.
I enjoy projects that combine creative problem-solving with disciplined engineering — especially those that transform complex ideas into real, measurable impact.
- Project Mesa https://github.com/projectmesa
Open Source Collaborator (Polars + Async Data Pipeline)
- Engineered a Polars-backed async data collector that scales to 1M+ agents
- Achieved 2–3× faster runtimes for agent-based simulations
- Transitioned backend from Pandas → Polars and modularized core APIs
- Designed a production-scale AI-Bidder system (custom prediction framework pivoted from RL)
- Built an LLM BI tool using RAG + RLHF
- Developed real-time inference flows and scalable ETL pipelines
- Implemented async logging, reproducibility workflows, and scalable data collection
- Optimized the simulation backend for large distributed runs