Purvag Patel

Using AI to rebuild local journalism in America, one community at a time. Building a platform that combines always-on data sensors, AI-driven curation, and agentic content generation to produce hyperlocal news at scale and fill the gap left by shrinking newsrooms.

Architected and built a 0-to-1 revenue-generating product, transforming the company into an AI-native animation studio. Led end-to-end development of Invisible Studio, a proprietary GenAI platform serving consumer and enterprise clients by automating content creation from script to high-fidelity video. Established a technical moat through multi-agent orchestration layers and IP-consistent fine-tuning pipelines using LoRA.

Established generative AI governance frameworks and led the data governance foundation across the organization.

Led the development of an end-to-end machine learning model deployment and hosting solution, boosting system reliability and developer productivity. Leveraged a diverse tech stack including AWS SageMaker, Kafka, Redis, ONNX, and TensorFlow to build scalable solutions, while fostering team growth and streamlined processes.

Led a cross-functional science and engineering team to optimize online ad auctions (Adtech) by building automation and scalable machine learning infrastructure powering a multi-billion dollar Ads program.

Managed a team enhancing Zillow’s home shopping experience with search services. Architected machine learning solutions for 15K QPS services, significantly improving personalized shopping metrics. Led a revamp of the search services and supported the microservice architecture handling 80% of company traffic.

As lead engineer, spearheaded the development of personalized home recommendation systems, collaborating closely with data scientists and machine learning engineers. Hands-on work included building regression models and scalable ML pipelines.