- Researched and integrated LLM models into existing product functionalities to enhance performance and user experience. - Orchestrated the AI model training, monitoring and re-train system for the entire Deep Learning life cycle. - Streamlined the AI model training process and reduced training time by 40%, resulting in faster time-to-market for AI-driven products. - Implemented a comprehensive labeling strategy, leading a team of 3 members to ensure adherence to labeling rules, resulting in a 30% improvement in data accuracy and a 20% reduction in labeling turnaround time. - Created 2 ETL pipelines with Dagster to generate aggregated tables for comprehensive monitoring. - Collaboratively developed and maintained 8+ weekly/monthly data analysis reports for 3 distinct teams. - Optimized the monthly data reports, ETL pipelines by reducing around 30% computing time. - Ingested data from disparate data sources using a combination of SQL, Google big query, AWS S3 bucket with Python.

- Authored an AI medical paper on corneal ulcer detection using Resnet, DenseNet, Inception_V3, and EfficientNet models, achieving 80% accuracy, enhanced 14% compared to doctors' diagnosis. - Achieved 93% accuracy in Pterygium disease classification using 6 CNN models and various data augmentation techniques.