- Developed AI recommendation system using TensorFlow and OpenAI CLIP, achieving 76% match accuracy.
- Designed a multi-database system (PostgreSQL, MongoDB) to handle 1TB+ of structured/unstructured data.
- Managed product documentation and workflows using SharePoint and collaboration tools.
- Developed real-time transcription (81% accuracy) using Whisper (OpenAI), processing 120+ requests daily.
- Implemented lecture summarization and detailed notes using Google Gemini LLM (88% accuracy).
- Integrated Clerk authentication for user sessions with 98% reliability.
Tech Stack: Next.js, Express, OpenAI, Gemini, AWS S3, Clerk
- Engineered a scalable audio processing pipeline, achieving 78% speaker identification accuracy with 100ms latency.
- Developed an efficient cloud storage architecture with AWS S3, reducing local infrastructure costs by 60%.
- Co-developed an AI-powered platform for meeting transcription and summarization, achieving 81% accuracy.
- Created an ETL pipeline processing 2.5M+ rows, reducing query time by 75% using BigQuery.
- Developed statistical models in R predicting user behavior with 89% accuracy.
- Built interactive Power BI dashboards for real-time KPIs, reducing reporting time by 85%.
Tech Stack: Python, OpenCV, MediaPipe, WebSocket
- Developed a real-time gesture recognition system with 74% accuracy using MediaPipe and custom CV algorithms.
- Implemented WebSocket connections for 60fps video processing with 16ms latency.
- Created an efficient caching system, reducing memory usage by 40% while maintaining 30fps performance.
