- Developed advanced PySpark scripts with AWS Glue to automate data validation and ETL processes, achieving a 25% reduction in manual effort and enhancing data accuracy and processing speed.- Architected and built event-driven data pipelines using AWS services (Kinesis Firehose, SNS, S3, SQS, Lambda, DynamoDB), enabling real-time data ingestion and improving overall processing efficiency.- Enhanced data pipeline reliability through complex SQL and database interaction modules, increasing data retrieval speed and seamless data flow.- Led the implementation of CI/CD automation with GitLab and AWS CDK, standardizing infrastructure management and accelerating deployment frequency by 40%.- Spearheaded a comprehensive onboarding program for new hires, mentoring 45+ employees on problem-solving and collaboration strategies, enhancing integration and strengthening team cohesion.- Served as a strategic internal business advisor for the development team on staffing module enhancements, improving operational productivity by 20% and optimizing resource allocation for greater project impact.

Enhanced YOLO and R-CNN models in PyTorch for large crowds, improving accuracy in occluded scenarios.Created PowerBI dashboards to visualize model performance, improving stakeholder communication and insights.Built LSTM model in Keras for time-series predictions, capturing temporal patterns and improving accuracy.Presented findings and insights, simplifying complex concepts to enhance collaboration and boost team productivity.

- Transformed large-scale biomedical PubMed data into vector representations using BioBERT pre-trained models. - Developed a semantic search API for Biomedical data and deployed Space Partition Tree and Graph (SPTAG) to construct indices based on the generated k-nearest vectors, enabling efficient and accurate search operations. - Conducted thorough analysis of search results' accuracy using Tableau, providing insightful reports and recommendations to the management team.