Siva Reddy

About Siva Reddy

As an AI/ML and Data Engineer with 5+ years of experience, I specialize in building…

Experience

1. Ai Ml Engineer

   QVC

   May 2024 — Present · PA, US

   Responsibilities: Architected and delivered a production-ready GenAI and Agentic AI platform leveraging GPT-4, LLaMA, LangChain, LangGraph, and AWS Bedrock, enabling real-time semantic product search, contextual Q&A, personalized recommendations, and automated customer service. Designed advanced RAG pipelines with Hugging Face Transformers, Pinecone, FAISS, and Bedrock-hosted foundation models, achieving 92% retrieval precision, reducing LLM inference latency by 25%, and increasing product discovery CTR by 14%. Engineered scalable vector database integrations (FAISS, Pinecone) and Bedrock API orchestration for multi-model selection and contextual reasoning. Developed agentic workflows using LangGraph and LangChain Agents, enabling autonomous reasoning, multi-step decision-making, and tool orchestration for complex customer queries. Fine-tuned GPT-4 and LLaMA models via LoRA/QLoRA and integrated Bedrock endpoints for secure and scalable inference across multiple LLMs. Applied prompt engineering techniques (dynamic templates, role conditioning, multi-shot query handling) for RAG and Agentic AI contexts. Deployed AWS SageMaker inference endpoints and Bedrock model APIs integrated with Redshift, Glue, and S3 to deliver real-time personalized analytics and customer segmentation. Automated model retraining and deployment workflows with AWS MWAA (Airflow), incorporating evaluation metrics (ROUGE, BLEU, cosine similarity) for continuous performance tracking. Established LLMOps observability using CloudWatch for latency, drift detection, and usage analytics, ensuring production reliability and scalability.

Education

• The University of Texas at Arlington

  Master's degree

• GITAM University

  Bachelor of Technology - BTech