At Apple, I build on-device machine learning pipelines and optimize transformer models for mobile deployment, achieving 40% energy savings, 2× faster inference, 50% smaller models, and a 60% speedup. I also develop multilingual NLP, wake-word detection, MLOps automation, and secure Generative AI integrations.
Previously at IBM and Accenture, I delivered production ML systems for petabyte-scale advertising data, real-time bidding, recommendation, ranking, fraud detection, and cloud deployment. I've improved CTR prediction, reduced inference latency and compute costs, and built reliable MLOps, monitoring, experimentation, and CI/CD workflows using Python, Spark, PyTorch, AWS, Kubernetes, and Docker.
