Varun M

Designed and deployed LLM-driven GenAI systems using Azure OpenAI, LangChain, LangGraph, CrewAI, and AutoGen to automate pharmaceutical workflows and clinical data processing. Built Retrieval-Augmented Generation (RAG) pipelines integrating Azure Cognitive Search, ChromaDB, and FAISS for regulatory document intelligence. Fine-tuned Large Language Models for clinical note summarization and drug safety analysis, reducing manual workload by 60%. Developed predictive ML models (XGBoost, TensorFlow) for demand forecasting and inventory optimization across global supply chains. Automated ETL and API pipelines with PySpark and FastAPI, integrating models into Azure ML and Databricks. Implemented AI observability using MLflow, LangSmith, and Azure Monitor to track hallucination rates and model accuracy.

Developed AI/ML models for predictive analytics, anomaly detection, and automation across Dell’s global infrastructure. Built scalable data pipelines and feature-engineering workflows using PySpark and Pandas for large-volume telemetry data. Designed hardware failure prediction and capacity planning models using TensorFlow, XGBoost, and Scikit-learn. Created and deployed REST APIs with Flask/FastAPI to serve model outputs to enterprise dashboards. Implemented RAG pipelines with LangChain, ChromaDB, and Azure Cognitive Search for knowledge retrieval and ticket summarization. Integrated LLM-based automation for report generation and incident summaries, reducing manual triage time by 45%.

Designed and implemented ML pipelines for anomaly detection, DDoS prediction, and network performance forecasting. Built time-series and deep learning models in Azure Databricks to detect latency spikes and traffic anomalies. Automated large-scale data preprocessing and feature engineering with PySpark and Pandas. Developed classification and anomaly-detection systems integrated into Cloudflare’s WAF and Bot Management platforms. Created ETL and inference APIs with Flask, FastAPI, and Azure Functions for real-time monitoring. Built dashboards with Power BI and Azure Monitor to visualize predictive insights and model reliability.

Built demand-forecasting and price-optimization models using regression and ensemble algorithms to increase sales efficiency. Automated data cleaning and transformation pipelines for large retail datasets using Pandas and NumPy. Created recommendation and customer-segmentation models to enhance loyalty-program engagement. Delivered data-driven insights to merchandising and logistics teams through interactive Power BI dashboards.

Designed predictive maintenance and anomaly-detection models to prevent outages and optimize bandwidth. Developed customer-retention and usage-segmentation algorithms using clustering and regression methods. Built performance dashboards (Tableau/Grafana) for churn tracking and service-level analytics. Integrated predictive insights into network monitoring and ticketing systems for proactive response.

Designed personalized recommendation engines and dynamic pricing models for e-commerce optimization. Built forecasting models (ARIMA, Prophet) for demand prediction and inventory management. Automated end-to-end ML pipelines with Airflow and AWS Lambda, supporting real-time retail insights. Deployed models on AWS SageMaker and EC2, monitored via MLflow and Power BI dashboards. Conducted A/B testing and uplift analysis to measure model performance and ROI.