work Financial Data Pipeline at JPMorgan Distributed Spark pipelines for multi-terabyte datasets research Amazon Nova Challenge Red-teaming LLM agents for code generation safety LLM Hallucination Detection Investigating factuality and attribution in large language models Knowledge Distillation for Reasoning Distilling chain-of-thought reasoning into smaller models opensource PyTorch Torchvision Contributions Open-source contributions to PyTorch computer vision library
