Bhumika M.

Bhumika M. is a skilled Machine Learning Engineer with expertise in Natural Language Processing (NLP) and Large Language Models (LLMs), specializing in building production-grade AI systems for information retrieval and applied AI. During her time at Fox Corporation, she developed sophisticated LLM-based metadata-extraction pipelines that enhanced content discovery and retrieval for millions of users by working with speech transcripts and catalog text. Her role involved tackling the entire model lifecycle from data preprocessing and fine-tuning to inference optimization and deployment, utilizing tools like Python, PyTorch, and Hugging Face. She demonstrated advanced skills in semantic embeddings and vector search, playing a crucial part in the media personalization and recommendation workflows on the Databricks platform. Additionally, her project work includes building an end-to-end retrieval pipeline with DocuMind and fine-tuning transformer models for news headline classification, achieving notable efficiency and accuracy improvements.