Siva Rama Krishna

Developed FastAPI-based microservices and REST APIs for customer onboarding, loan origination, and AI-powered underwriting workflows, reducing application response times by 35%. Built credit risk ML models with scikit-learn, improving loan risk classification accuracy by 18%. Implemented RAG solutions using OpenAI GPT-4, LangChain, Pinecone, and prompt engineering for policy-aware credit recommendations and document summarization. Created LLM-powered document intelligence services using Hugging Face Transformers, Python, and PyTorch to extract borrower and compliance data from unstructured documents. Integrated enterprise banking systems, credit bureaus, and repositories via REST APIs, PostgreSQL/SQLAlchemy, and Apache Kafka for real-time workflow data. Deployed microservices on AWS (Docker/Kubernetes) with Terraform and automated CI/CD using GitHub Actions; contributed to PyTest coverage, code reviews, and secure, compliant delivery in Agile.

Architected Django and FastAPI microservices with REST APIs, OAuth2/JWT, and RBAC to support secure multi-tenant clinical trial platforms. Built React/TypeScript dashboards to improve researcher workflows for patient search, eligibility visualization, and trial management. Developed NLP pipelines using PyTorch, Hugging Face Transformers, and spaCy to extract ICD codes, medications, and clinical entities from 500K+ unstructured documents. Improved patient eligibility prediction with scikit-learn models combining structured patient records and NLP-extracted features for better trial candidate ranking. Implemented Elasticsearch for full-text search and PostgreSQL for high-performance querying across 2M+ patient records. Deployed containerized services on AWS (EC2/ECS/S3/Lambda), automated CI/CD with Jenkins, and supported reliable production releases.

Built Django-based microservices and REST APIs (Django REST Framework) for product catalog, shopping cart, order management, and payment workflows. Developed scikit-learn recommendation models using Python, Pandas, and NumPy to personalize suggestions based on purchase history and browsing behavior. Created XGBoost forecasting models for inventory planning and replenishment using historical sales, inventory, and seasonal trends. Optimized MySQL queries and Django ORM performance to reduce response times for search, checkout, and order processing. Strengthened security and quality through RBAC, unit/integration testing, and code reviews. Deployed containerized services on AWS (EC2/ECS/RDS/S3) with Docker and automated CI/CD pipelines using Jenkins and Git for faster releases and higher availability.