Uday-- kurella

At Discover Financial Services, I design and deliver enterprise conversational AI and intelligent automation solutions, supporting customer engagement, credit risk analysis, fraud detection, and operational workflows in a highly regulated financial environment. I lead the development of chatbots and virtual assistants using Kore.ai, building intent-driven conversational flows, dialog orchestration, entity extraction, and backend integrations to automate call-center and servicing use cases. These solutions enable context-aware, scalable conversational experiences that improve response times and customer satisfaction. I design cloud-native AI services aligned with AWS architectures, integrating conversational AI platforms with real-time ML inference services for credit decisioning and fraud prevention. I apply AI agent–based approaches that combine conversational interfaces with automation logic and enterprise system integrations to execute end-to-end workflows such as document intake, loan processing, and customer query resolution. I leverage Generative AI techniques, including large language models, prompt engineering, and Retrieval-Augmented Generation (RAG), to enhance document understanding and automate processing of financial statements and loan applications. These capabilities significantly reduce manual effort and improve operational efficiency. To support Enterprise Data Management (EDM), I work closely with data engineering teams to integrate structured and unstructured financial data sources, standardize data access, and ensure governed inputs for conversational AI systems. I develop secure, scalable APIs and deploy containerized services to enable reliable integration across enterprise platforms.

At Hackensack Meridian Health, I designed and delivered enterprise conversational AI and intelligent automation solutions to support clinical operations, patient engagement, and hospital resource optimization within a highly regulated healthcare environment. I collaborated with data scientists, clinicians, and operational teams to build AI-driven automation and decision-support systems, translating complex clinical workflows into structured, scalable AI solutions. My work focused on enabling natural language–driven interactions, data-driven insights, and real-time automation to improve patient outcomes and reduce hospital readmissions. I developed NLP-powered conversational and text analytics pipelines using Python, SpaCy, and transformer models to extract insights from unstructured clinical notes and EHR data. These capabilities supported AI-assisted workflows for diagnostics, patient risk stratification, and operational planning, forming the foundation for intelligent assistant and agent-based use cases. I designed cloud-native ML pipelines aligned with enterprise and AWS-style architectures, ensuring secure, scalable deployment of predictive services and automation workflows. I worked extensively with Enterprise Data Management (EDM) concepts, integrating structured and unstructured clinical data while enforcing governance, access controls, and interoperability standards such as FHIR. I implemented AI automation pipelines with containerized services, enabling reliable integration between AI models, data platforms, and downstream clinical systems. My work included developing time-series forecasting models to predict ER visit volumes, supporting proactive staffing and resource allocation. To ensure trust and compliance, I applied explainable AI techniques to support transparency in clinical decision-making and auditing. I partnered closely with cybersecurity teams to enforce data security, encryption, and privacy controls.

At Brown & Brown, I led end-to-end data science projects to support risk modeling, pricing optimization, fraud detection, and customer insights across multiple insurance lines. I designed and deployed scalable ML pipelines using AWS services (S3, Lambda, Glue, SageMaker), enabling secure, cost-effective predictive applications. One of my key contributions was developing real-time quote and customer segmentation APIs powered by behavioral and historical data, driving more accurate policy recommendations. I applied advanced NLP techniques to analyze customer feedback, underwriting notes, and call transcripts—enhancing customer satisfaction metrics and risk profiling. I implemented time series forecasting models for claims prediction and staffing optimization, improving operational planning and budgeting accuracy. Using deep learning models (PyTorch, TensorFlow), I improved fraud detection precision, contributing to significant cost savings. I also championed model governance using DVC and MLflow, ensuring full traceability and compliance in a highly regulated environment. Additionally, I supported the cloud migration of legacy data science workflows, enhancing scalability, collaboration, and automation. To visualize key metrics such as loss ratios and retention rates, I built interactive dashboards using Tableau and Plotly Dash. I conducted A/B testing and uplift modeling for marketing and retention strategies, directly influencing improved campaign ROI. I mentored junior data scientists and helped instill strong MLOps practices, working closely with DevOps to build CI/CD pipelines for ML deployments. Throughout, I maintained detailed documentation and enforced data privacy policies (HIPAA, SOC 2) in coordination with cybersecurity teams.

As a Data Science Engineer at AutoZone, I led the development and deployment of predictive analytics solutions that significantly improved retail operations across 400+ stores. A key project involved building ML models to optimize inventory management, reducing stockouts by 18% and enhancing supply chain efficiency. I designed scalable ETL pipelines using AWS Glue, Redshift, and S3 to process high-volume transactional data. Leveraging Amazon SageMaker, I deployed machine learning models including Random Forest, Gradient Boosting, and K-Means Clustering to analyze customer behavior and power product recommendations. I created NLP tools using Python (spaCy, NLTK) to perform sentiment analysis on survey and social media feedback, driving actionable service improvements. I also applied time series models like ARIMA, Prophet, and LSTM for seasonal sales forecasting and demand planning. Collaborating with DevOps, I implemented CI/CD pipelines using Docker and AWS CodePipeline for model deployment, improving version control and deployment cycles. I led A/B testing and uplift modeling for targeted marketing campaigns, which boosted ROI by 12%. To visualize insights, I developed dashboards in Tableau and Power BI, presenting KPIs, sales trends, and customer segments to leadership. I also engineered solutions to unify data from POS systems, CRMs, and APIs, enabling a 360° customer view. My role included maintaining cloud infrastructure using Lambda, CloudFormation, and EC2, and contributing to data governance initiatives around quality, security, and versioning. I worked closely with cross-functional agile teams to deliver data science solutions aligned with key retail strategies.

As a Data Engineer at CenturyLink, I designed and maintained large-scale ETL pipelines using Python, Spark, and SQL to process telecom data from diverse sources. I built a scalable data lake on Amazon S3, which significantly improved data accessibility and reduced downstream query latency for analytics teams. I engineered event-driven data pipelines using AWS Glue and Lambda to automate processing and reduce data delivery delays. Using Amazon Kinesis, I enabled real-time streaming of Call Detail Records (CDRs), which improved fraud detection and operational responsiveness. A key project involved optimizing Redshift data warehouses through advanced partitioning and indexing strategies, resulting in faster report generation and reduced costs. I also automated data quality validation using PySpark and Airflow, decreasing data integrity issues by 40%. To support infrastructure as code, I deployed and maintained Terraform scripts for consistent and secure AWS provisioning. I led efforts in metadata management using AWS Glue Data Catalog and collaborated with security teams to enforce IAM roles, encryption standards, and VPC isolation, ensuring compliance with telecom industry regulations. Additionally, I developed dashboards using Tableau and QuickSight for real-time monitoring of network performance, enabling business teams to make faster decisions. I supported cloud migration projects and mentored junior engineers on best practices in data engineering, AWS architecture, and version control using Git.

Pursuing a Master’s degree in Data Science with a strong focus on machine learning, deep learning, natural language processing, and scalable data systems. Coursework includes Statistical Learning, Big Data Analytics, Deep Learning, and AI Ethics. Engaged in hands-on projects involving real-world datasets and cloud computing. Collaborated with peers on Kaggle-style competitions and participated in AI/ML workshops, contributing to both academic and industry-relevant research.