Summary
John D'souza is a Director of Machine Learning based in San Francisco with 11 years of experience building production-grade NLP and generative AI systems that have driven over $100M in lifetime value. He has led end-to-end Document Intelligence and QA initiatives—combining fine-tuned LLMs, prompt engineering, and MLOps—to deliver $35M in revenue, boost student engagement by 20%, and expand a document library by 40%. At Course Hero and Altana he established repeatable practices for data management, model training, deployment, and observability while growing and mentoring a cross-functional team of engineers. Technically fluent in PyTorch, TensorFlow, Keras and modern NLP techniques, he pairs deep engineering craft with product-driven roadmaps to align ML work to measurable business outcomes. An uncommon strength is his track record of turning research-grade models into scalable, revenue-generating features through pragmatic platformization.
11 years of coding experience
18 years of employment as a software developer
B.E Electronics and Communication, B.E Electronics and Communication at Visvesvaraya Technological University