Summary
Joshua Shapiro is a Staff Machine Learning Engineer with seven years of experience building production-grade ML systems, currently leading development of AKASA's medical coding solution and inference engine for LLMs exceeding 10B parameters. He combines research-to-production expertise—from language modeling and conversational AI at ASAPP to deep learning and temporal state detection at IBM Watson Research—with a focus on scalable, compute-agnostic tooling that accelerates iteration and cuts latency and training time. His recent work reduced inference latency 7x using VLLM, lifted entity extraction accuracy by 20%, and standardized offline evaluation to closely mirror production outcomes. Joshua also teaches a senior capstone course at George Washington University, bringing industry best practices into the classroom and mentoring the next generation of engineers. Comfortable across NLP, vision, and systems, he’s notable for automating heterogeneous compute workflows and shipping infrastructure that makes large models practical in regulated healthcare settings.
7 years of coding experience
11 years of employment as a software developer
Bachelor’s Degree Computer Science, Bachelor’s Degree Computer Science at The George Washington University - School of Engineering & Applied Science
Bachelor’s Degree Computer Science, Bachelor’s Degree Computer Science at Korea University
High School, High School at Jesuit High School