Kiran Teja Sriperumbuduri

About Kiran Teja Sriperumbuduri

AI/ML Engineer with 3+ years of experience building scalable machine learning and Generative AI solutions. Skilled in LLMs, RAG pipelines, and end-to-end ML systems, with hands-on expertise in AWS, Azure, and MLOps practices. I focus on developing production-ready AI applications that improve decision-making, enhance efficiency, and deliver real business value. Passionate about Generative AI, agentic workflows, and deploying intelligent systems at scale.

Experience

1. Machine Learning Engineer

   HCLTech

   Dec 2025 — Present · OH, US

   Designed and deployed end-to-end ML pipelines on Azure Databricks and Azure ML, processing data from 25+ sources to build risk scoring models for credit assessment and fraud detection workflows. • Built scalable feature engineering pipelines to process 15M+ records daily, reducing feature preparation cycles and enabling faster model training and retraining workflows. • Engineered LLM-powered RAG systems using LangChain and Azure OpenAI, integrating retrieval from Snowflake vector stores to enhance fraud detection queries with contextual, real-time insights—boosting response accuracy by 30% for investigative workflows. • Fine-tuned domain-specific LLMs (Llama 3, Mistral) for credit risk explainability and anomaly summarization, deploying RAG pipelines on Kubernetes to deliver interpretable predictions and reduce manual review time by 40%. • Developed and tuned supervised learning models (XGBoost, Scikit-learn, PyTorch) for credit risk prediction and anomaly detection, improving model accuracy and prediction reliability across datasets. • Implemented near real-time inference pipelines using Azure Event Hubs, Kafka, and REST APIs, enabling low-latency scoring for high-volume transaction streams. • Integrated Snowflake and Azure Lakehouse environments to create governed, ML-ready datasets and feature stores, supporting reproducible experimentation and model consistency. • Containerized ML services using Docker and deployed on Kubernetes, managing multi-instance model serving environments for scalable and reliable production workloads. • Established MLOps monitoring frameworks to track model drift, data quality, and inference performance, reducing production issues and improving system observability. • Collaborated with cross-functional teams to translate business requirements into deployable ML solutions, accelerating model deployment cycles and increasing adoption across analytics and business teams.

Education

• Anderson University (SC)

  Master's degree, Computer Science

• Jawaharlal Nehru Technological University Kakinada (JNTUK)

  Bachelor of Technology - BTech, Electronics & Communication Engineering