# Stephen O'Sullivan > SWE | ML | MIT PhD Location: San Francisco Bay Area, United States Profile: https://flows.cv/stephenosullivan PHYSICS PHD (MIT) → ML SYSTEMS & PHYSICAL AI ENGINEER I design, build, and deploy machine learning systems that solve ambiguous, high-impact business problems, especially where digital intelligence meets the physical world. From gravitational wave modeling and numerical simulations to production ML pipelines in retail, fintech, and robotics, I bridge theoretical insight with practical, real-time delivery. ​ Recent work includes model predictive control and compliant motion for collaborative robot manipulators in automotive assembly, diagnostic ML for in-store retail operations (causal modeling, anomaly detection, and forecasting), and end-to-end credit risk and pricing models for small business lending. I enjoy taking messy, high-dimensional time-series and event data, turning it into robust models, and hardening those models for production. ​ TECHNICAL FOCUS: ML systems and MLOps (Docker, Kubernetes, Airflow), deep learning in PyTorch/TensorFlow, probabilistic modeling and graphical models, time-series analysis (forecasting, anomaly detection), and numerical methods for complex physical systems. ## Work Experience ### Principal Software Engineer @ Ghost Autonomy Jan 2022 – Jan 2023 | Mountain View, California, United States - Characterized advanced driver assistance control performance as a function of sensor latency and noise, using simulation and analysis of time-series data from perception and control stacks. ​- Quantified robustness and failure modes of control algorithms, informing requirements for sensor quality, timing, and system-level safety. ### Principal Software Engineer @ Symbio Robotics Jan 2021 – Jan 2022 | Emeryville, California, United States Developed control and planning algorithms for collaborative robot manipulators in automotive assembly, with a focus on compliant motion and safe human–robot interaction. Applied model predictive control (MPC) and trajectory optimization to achieve robust, high-throughput assembly tasks under uncertainty. ​ Led architecture refactoring of core robot software to improve reliability, debuggability, and integration of advanced control algorithms into production systems. ### Machine Learning Lead @ Walmart Global Tech Jan 2020 – Jan 2021 | San Francisco Bay Area Led diagnostic ML modeling for in-store activity using transactional and sensor-derived time-series data, focusing on causal analysis, anomaly detection, and forecasting to improve operations. ​ Designed and deployed production pipelines for activity monitoring and forecasting, collaborating closely with data engineering and product teams. ​ Matured MLOps practices across a distributed data science team, including standardization of model deployment, monitoring, and experiment tracking ### Senior Machine Learning Engineer @ Kabbage from American Express Jan 2017 – Jan 2020 | San Francisco Bay Area Upgraded bank transaction parsing from rules-based features to embedding-based representations using NLP techniques, improving fidelity of customer banking history modeling. ​ Built and operationalized probabilistic graphical models for automated pricing optimization of credit lines and interest rates. ​ Mentored an intern in graph-based fraud detection using customer network segmentation and link analysis. ​ Delivered a European credit risk model for ING Bank via transfer learning from U.S. data, leading the full lifecycle from stakeholder alignment through deployment into production. ### Automated Driving Research Intern – Machine Learning Embedded Software @ Ford Motor Company Jan 2017 – Jan 2017 | Palo Alto Designed a procedural system for dynamic simulated driving environments with virtual driver-in-the-loop control using Unreal Engine, supporting research in automated driving. ## Education ### Doctor of Philosophy (Ph.D.) in Physics Massachusetts Institute of Technology Jan 2006 – Jan 2015 ### Research Assistant in Physics University College Cork Jan 2004 – Jan 2006 ### Bachelor's degree in Theoretical and Mathematical Physics Trinity College Dublin Jan 2000 – Jan 2004 ## Contact & Social - LinkedIn: https://linkedin.com/in/stephengosullivan - GitHub: https://github.com/stephenosullivan - Website: http://gmunu.mit.edu/viz/embed_viz/embed_viz.html --- Source: https://flows.cv/stephenosullivan JSON Resume: https://flows.cv/stephenosullivan/resume.json Last updated: 2026-03-23