JPMorgan Chase&co
|AI/ML Engineer
Alabama, Alabama, US
Summary
Leads the development and deployment of enterprise-scale AI/ML and deep learning solutions, optimizing data pipelines and enhancing model performance for critical financial and legal applications.
Highlights
Curated and preprocessed over 5M internal documents using Python (pandas, NumPy, spaCy, NLTK), improving input quality and reducing irrelevant model outputs by 25%.
Built and fine-tuned PyTorch-based LLMs using Hugging Face Transformers, reducing manual content preparation time by 35% for 60,000+ employees via real-time content generation.
Integrated FAISS-based retrieval-augmented generation (RAG) with LLM pipelines, increasing factual accuracy of outputs by 20% through relevant internal document referencing.
Optimized GPU cluster training pipelines on AWS with incremental retraining strategies, reducing model training time by 40% and enabling weekly model updates.
Led A/B testing of model prompts and retrieval strategies, achieving a 25% reduction in manual editing and improved trust in AI-generated content.