Senior Data Scientist / AI Engineer at Wipro

Led integration and fine-tuning of large language models (LLMs) and multi-modal vision-language models (VLMs) applied to clinical record summarization and medical imaging reports, optimizing inference and deployment on Google Cloud Platform (GCP) using ONNX, TensorRT, and NVIDIA Triton for GPU acceleration. Developed conversational AI agents and Gen AI workflows leveraging technologies such as LangChain, Hugging Face, and OpenAI API. Engineered forecasting and feature engineering pipelines supporting cognitive and behavioral medical analytics. Built scalable real-time streaming and ETL pipelines integrating BigQuery, Firestore, and Apache Spark. Created Power BI dashboards to enable actionable clinical insights and applied AI governance frameworks for compliance and security.

Data Scientist at Fidelity Investments

Developed AI-driven forecasting models improving regional demand prediction for pharmaceutical supply chains deployed on AWS. Designed robust ETL pipelines and real-time streaming applications using AWS Glue, Lambda, Kinesis, and SNS. Applied deep learning computer vision models using TensorFlow and Keras to enhance defect detection accuracy. Implemented NLP models and RAG pipelines for customer sentiment and multi-modal analysis. Optimized ML model inference latency with ONNX and TensorRT deployed using NVIDIA Triton. Supported AWS migration of ML workflows with CI/CD and MLOps best practices. Created interactive dashboards for KPIs and business intelligence leveraging Power BI and Django.

Data Scientist at Thomson Reuters

Built classification pipelines and forecasting models using machine learning techniques such as Random Forest, Logistic Regression, ARIMA, and LSTM to improve warehouse logistics and supply chain demand prediction. Designed and deployed large-scale ETL pipelines with Azure Data Factory and Databricks. Created unified feature stores to enhance model reproducibility, and automated model retraining with Azure ML Pipelines. Performed model explainability analysis with SHAP and LIME, and optimized hyperparameters with GridSearchCV and Bayesian Optimization. Delivered interactive Tableau dashboards for monitoring model and business performance.

Data Analyst at Reliance Retail

Developed interactive Tableau dashboards and SQL datasets tracking sales, stock, and supplier performance. Automated reporting workflows using Python and Excel to improve data validation and audit processes. Collaborated with regional teams to define KPIs and reporting requirements. Conducted root cause analysis to resolve inventory and pricing discrepancies. Maintained data dictionaries and product taxonomy mappings to ensure consistent reporting across retail operations. Documented standard operating procedures to improve transparency and knowledge transfer.

Senior Data Scientist/AI Engineer at Wipro

Engineered multi-agent orchestration using LangChain, LangGraph, MCP, and DeepSpeed to automate compliance and marketing workflows, improving accuracy and reducing manual escalations. Built retrieval pipelines with LlamaIndex, FAISS, Pinecone, and hybrid RAG approaches, boosting structured and unstructured intelligence accuracy. Directed MLOps and production deployments across Azure ML, SageMaker, Bedrock, Vertex AI, and GCP with CI/CD, Docker, Kubernetes, Kubeflow, and drift monitoring to reduce latency and improve reliability. Implemented prompt engineering and fine-tuning across GPT and open models, and enabled human-in-the-loop validation and safety checks. Deployed GenAI apps with AWS Bedrock, Lambda, S3, Step Functions, DynamoDB, and Kubernetes-based microservices. Orchestrated multi-agent pipelines with evaluators, memory-aware reasoning, and feedback loops to reduce manual review escalations. Automated NLP pipelines for classification, extraction, and research synthesis via Pyt

Data Scientist at Fidelity Investments

Built time-series models (ARIMA, Prophet, LSTM) to forecast NAV fluctuations and redemption volumes, enabling stronger liquidity planning. Optimized XGBoost and scikit-learn models for churn and fraud detection with feature engineering on transactional and behavioral data. Implemented custom model evaluation, fine-tuning pipelines, and PEFT-based domain adaptation for enterprise GenAI optimization. Designed anomaly detection workflows (TensorFlow/Keras) and deployed APIs with FastAPI and CI/CD. Developed RFM-based segmentation to reduce churn by 12%. Built scalable ETL pipelines with AWS Glue, Lambda, and Spark to automate retraining, cutting manual data prep by ~30%. Created NLP pipelines (spaCy/NLTK) to extract sentiment from analyst reports and call-center transcripts, improving client satisfaction. Delivered dashboards in Power BI/Matplotlib to monitor risk and exposure. Migrated legacy ML workflows to AWS with containerized deployments and CI/CD. Built SHAP/LIME based audit and co

Data Scientist at Thomson Reuters

Implemented a legal analytics platform using logistic regression and decision trees to flag low-engagement content, improving editorial prioritization by 25%. Designed SQL/NoSQL data lakes to support modeling and clustering of legal user behavior. Conducted EDA on access logs to inform taxonomy design. Built Tableau dashboards to visualize predictive scores, churn, and citation trends. Automated weekly reporting with Python, reducing manual reporting time by 30%. Developed secure ML pipelines with Azure Synapse and Blob Storage, ensuring versioned embeddings and reliable retraining for classification tasks. Applied PCA/t-SNE for dimensionality reduction to improve content taxonomy. Built gradient-boosting models to forecast licenses, increasing accuracy by 18% and streamlining knowledge-base design. Evaluated models using precision/recall/RMSE.