Constantly learning and exploring, diving deep into research papers, ML framework documentation, and source code to build efficient ML pipelines. Staying up to date with the latest trends in machine learning and AI to drive impactful innovation.• Utilized parallel processing and implemented distributed embedding techniques on large clinical notes (100K+ tokens) using pretrained clinical models based on BERT, RoBERTa, Longformer, GPT, etc. transformer architectures, resulting in over 3 times improvement in embedding speeds.• Utilized PyTorch Lightning and PEFT techniques (LoRA) to train transformer models with over 3 billion parameters on multiple A100 GPUs, resulting in efficient distributed training and scalability for large model development.• Implemented a cross-modal transformer model capable of cross-attention in PyTorch and performed hyperparameter tuning, resulting in state-of-the-art performance and contributing to 4 research papers and a journal submission.

• Researched and gathered time series data related to EC2 systems on AWS for optimizing a machine learning model, contributing to building a product that cut job delays on on-premise servers by 80%.• Developed over 50 REST APIs using Django and Spring Boot for an internal website, enhancing communication between systems and gaining proficiency in new frameworks and technologies.