Piyush Patil

Spearheaded end-to-end development for three flagship products (Notation AI, Scrubber, Fownd Website) leveraging React, Go, Electron, and FastAPI, accelerating high-quality product releases in a dynamic startup environment.

Designed and implemented a scalable multi-service microservices architecture for the Notation AI product, enabling real-time audio/video streaming (WebSocket, WebRTC) and managing data-intensive workflows with AWS S3 and SQS.

Built a robust, production-ready React frontend with seamless WebSocket integration for real-time audio transcription and dynamic updates, displaying ML-driven "Subjective" content.

Engineered advanced video processing pipelines by integrating a PoseNet bounding box model into a custom Go WebRTC server, facilitating efficient transmission of video streams and keypoint data to AWS S3 and SQS for ML processing.

Spearheaded cloud-native deployments using Docker, ALB for HTTP, NLB for UDP streaming, and EC2 auto-scaling, significantly improving system reliability and throughput.

Developed an interactive web platform using React, TypeScript, and Map Box, leveraging RESTful endpoints to enable seamless retrieval of over 150,000 road segment points.

Crafted a scalable backend utilizing AWS Lambda for automated fetching of spatial data from AWS RDS and external sources, delivering valuable spatial insights to users.

Designed robust REST APIs and developed Python-based data crawlers for AWS, automating data updates via CRON jobs; utilized data visualizations to identify significant trends and uncover high emission sources.

Established and maintained automated CI/CD pipelines with GitHub Actions and AWS, reducing production deployment time by 70% and enhancing testing infrastructure reliability.

Conducted advanced data querying and exploratory analysis on large datasets from Amazon RDS (PostgreSQL), collaborating effectively with cross-functional teams.

Utilized Tableau to create visually compelling dashboards and developed an automated Metabase dashboard and interactive website, converting complex air pollution data into actionable insights for stakeholders.