Lahari Katta

About Lahari Katta

Barclays leverages their expertise in Graph Neural Networks (GNNs), PyTorch Geometric…

Experience

1. Ai Ml Engineer

   Barclays

   Apr 2024 — Present · San Francisco, CA, US

   Contributed to building a predictive maintenance system for critical banking infrastructure using Graph Neural Networks (GNNs) implemented with PyTorch Geometric, assisting in model training and validation on AWS SageMaker.• Developed and maintained real-time data pipelines using AWS Glue, Apache Kafka, and Spark Structured Streaming to process sensor and operational data into graph-based inputs for models.• Explored LLM frameworks (Hugging Face, LangChain) for potential applications in anomaly detection and achieving 12% higher precision in log analysis, aligning with industry trends in generative AI.• Integrated MLflow for experiment tracking and reducing model versioning errors in CI/CD pipelines by30%.• Collaborated with senior engineers to optimize model hyperparameters and feature engineering, achieving significant improvements in predictive accuracy and early failure detection.• Maintained CI/CD pipelines using Docker, AWS CodePipeline, and ECS Fargate to automate testing, deployment, and scaling of ML models.• Ensured compliance with GDPR & banking regulations, passing 100% of internal audits for AI model governance.• Developed and executed rigorous testing procedures and ensured compliance with industry standards and regulatory requirements for system reliability and security.