Sriram Degala

About Sriram Degala

AI/ML Engineer with 4+ years of experience designing and deploying intelligent systems…

Experience

1. Ai Ml Engineer

   UT MD Anderson

   Aug 2024 — Present · US

   Built predictive cancer risk scoring pipelines using LangChain and LLaMA transformers, optimizing model architectures and training workflows to improve early-stage tumor detection accuracy by 27%.• Developed multimodal triage agents with GPT-4, Prompt Engineering, and RAG, integrating structured/unstructured oncology data for faster model-driven patient assessments and reducing query latency by 34%.• Deployed and managed ML models for imaging diagnostics using Kubernetes, Kubeflow, and automated CI/CD pipelines, enabling reproducible real-time inferencing at scale.• Automated longitudinal training pipelines on AWS EC2 with Hugging Face Transformers, handling continuous ingestion of oncology EHRs and retraining models with minimal manual intervention.• Engineered distributed imaging data workflows with Apache Beam and Google Dataflow, processing 100+ TB of radiology scans with optimized memory allocation, parallel scheduling, and data preprocessing.• Designed explainable ML dashboards with Dash (Plotly) and custom APIs, exposing model predictions, feature attributions, and performance metrics to oncology researchers and clinical data teams.• Automated experiment tracking and statistical evaluation of A/B tests using MLflow and custom Python workflows, ensuringreproducible comparisons and reducing model selection time by 20%.

Education

• Stevens Institute of Technology

  Master of Science - MS, Machine Learning

• Jaypee Institute Of Information Technology

  B.Tech, Computer Science and Engineering

  2018 — 2022