Raghu Ram Sai Ammula

Built and fine‑tuned LLM‑driven recommendation systems (GPT‑4 via Azure OpenAI, Llama 3 with LoRA/QLoRA) to deliver hyper‑personalized product experiences; increased engagement by 37% and boosted conversion across e‑commerce storefronts. Architected end‑to‑end RAG pipelines on Azure (LangChain + FAISS + Azure Cognitive Search), integrating multi‑source embeddings and context re‑ranking; improved semantic search accuracy by 42% and reduced irrelevant matches. Operationalized LLM workflows on Azure ML, Azure Functions, and AKS with autoscaling, containerized deployment, and Redis caching; cut inference latency by 31% and achieved 99.9% uptime for real‑time recommendations. Orchestrated model monitoring and retraining via Azure ML Pipelines, MLflow, and Application Insights dashboards; automated drift detection and pipeline triggers, sustaining >95% response faithfulness during traffic spikes. Developed and deployed Generative AI chatbots with FastAPI + Azure OpenAI AP

Developed and deployed large-scale recommendation systems via collaborative filtering and gradient boosting on Amazon SageMaker; boosted CTR by 21% and engagement across e‑commerce platforms. Designed distributed ETL pipelines with AWS Glue, Apache Spark, and Airflow to process multi‑terabyte user interactions; achieved 99.8% ingestion accuracy and reduced data latency from hours to minutes. Implemented NLP models (BERT, T5) for semantic search and product categorization; improved search relevance by 28% and enabled contextual product discovery for thousands of SKUs. Designed and deployed real-time fraud detection pipelines with Amazon Kinesis and anomaly detection; reduced false positives by 22% and prevented major financial exposure across transactional streams. Automated the end-to-end ML lifecycle with MLflow, Docker, and AWS Lambda for model versioning, evaluation, and retraining; integrated CI/CD and observability via CloudWatch, S3 versioning, and metric dashboards, achievin