Teja Chakilam

Built and optimized ETL and API-based data pipelines using Databricks, SQL, and Python across AWS and Azure to prepare, transform, and integrate enterprise datasets for analytics, predictive modeling, and Generative AI. Developed an enterprise GenAI Sales Assistant using LangChain and LangGraph with LLMs to generate personalized sales content and solution recommendations. Designed and evaluated RAG pipelines (embeddings, chunking, metadata filtering, and evaluation) to improve grounding/factual accuracy and stakeholder acceptance; analyzed retrieval quality and agent behaviors to reduce routing and fallback errors and improve end-to-end latency. Troubleshot production issues (context-window, retrieval, API failures, downstream errors) to support 99.9% availability at launch.

Developed Generative AI applications for banking operations, including automated document summarization and employee access to internal knowledge with actionable, data-driven insights. Built RAG pipelines connecting LLMs with approved banking documents and enterprise knowledge sources for context-aware responses. Used Python and SQL to analyze datasets and LLM evaluation results to identify retrieval gaps and response-quality issues; refined prompt templates and evaluated outputs for factual accuracy, relevance, consistency, retrieval quality, and latency. Integrated GenAI solutions with internal APIs and enterprise data sources while implementing access controls and safeguards for sensitive information.

Analyzed healthcare and pharmacy datasets with Python and SQL to identify utilization trends, cost drivers, and operational patterns for care management and pharmacy operations. Developed and validated predictive models using feature engineering, cross-validation, and statistical analysis to support risk assessment and utilization forecasting. Performed exploratory data analysis and data quality validation across disparate systems, identifying anomalies and integration gaps, and translating findings into recommendations. Built and optimized ETL pipelines, data warehouse workflows, and semantic data models using IBM DataStage, SQL, and relational databases for reliable reporting and analytics.