Passionate about Machine Learning, Computer Vision, Generative AI, and NLP with a robust academic background and professional experience spanning various domains of artificial intelligence. Building intelligent systems at scale.
Built and productionized large-scale Whisper ASR + translation pipelines for multilingual subtitles across EN/DE/ES, reducing vendor costs by $10,000+/month with fully automated training, evaluation, and deployment cycles via Airflow.
Designed and deployed real-time ASL to text translation system using MediaPipe segmentation and CNN-GRU architecture. Improved translation accuracy for 500+ users in the deaf-blind community with 3x interaction efficiency gains and seamless robot integration.
Developed state-of-the-art Vision Transformer recommendation system integrated with GPT-4 and BERT. Captured fine-grained visual features and correlated with textual data, achieving 22% CTR improvement and enhanced product discovery.