As a Software Engineer who builds AI systems that solve real problems at scale, I led a team at Teladoc Inc. that delivered an enterprise AI Copilot helping 500+ employees find answers instantly — cutting resolution time by 87%.
Experience
2025 — Now
2025 — Now
San Francisco, CA
• Designed and implemented a Generative AI workflow using LLMs and LangChain to automate structured comparison of financial disclosures across audit engagements, improving document review efficiency by 36% while maintaining traceable output logs for compliance validation.
• Built an NLP-based risk classification pipeline using Hugging Face Transformers to evaluate regulatory filings against predefined risk indicators, increasing early-stage risk detection accuracy by 29% during advisory assessments.
• Developed a retrieval-augmented internal knowledge assistant with refined prompt logic and context-window management, ensuring consistent, policy-aligned responses for tax and compliance consulting teams.
• Standardized model deployment by containerizing FastAPI inference services and orchestrating them on AWS with Kubernetes, introducing structured version control and monitoring aligned with internal AI governance standards.
• Engineered Spark-based preprocessing pipelines in Databricks to normalize large financial datasets prior to anomaly detection modeling, reducing data-quality related reprocessing efforts across advisory workflows.
• Implemented MLflow experiment tracking and performance monitoring to improve reproducibility, streamline model audits, and simplify documentation during internal compliance reviews.
• Partnered with risk analysts and data governance teams to translate model outputs into structured reporting artifacts, ensuring business stakeholders understood model assumptions, confidence levels, and operational limitations.
2023 — 2025
2023 — 2025
Dallas, TX
• Architecting and deploying enterprise-grade Generative AI solutions to transform employee productivity and information access. Built an internal AI Copilot using Python, Azure Cognitive Search, and Large Language Models (LLMs) serving 10,000+ employees, enabling conversational document retrieval and policy queries with 95% answer accuracy.
• Key contributions include developing Retrieval-Augmented Generation (RAG) pipelines that improved answer relevance by 60%, fine-tuning LLMs for contextual accuracy and factual grounding, and implementing prompt optimization strategies. Built Flask REST APIs to serve AI models with sub-150ms response times, integrating Azure Cognitive Search for semantic document retrieval with metadata filtering and secure access controls.
• Automated SharePoint API workflows using Python and Power Automate, reducing manual HR query processing time by 40%. Collaborate closely with Data Science, IT, and business stakeholders to identify AI use cases, integrate models into production workflows, and measure impact through user analytics and feedback loops.
• Ensure HIPAA compliance and data security in AI implementations while continuously optimizing model performance, inference latency, and system reliability. Champion responsible AI practices including bias detection, explainability, and human-in-the-loop validation for healthcare applications.
2021 — 2023
2021 — 2023
Hyderabad
• Built and maintained full-stack web applications using React and Python (Flask/Django), integrating ML-powered features including intelligent search ranking, anomaly detection dashboards, and automated content classification for enterprise clients.
• Developed and deployed Java Spring Boot microservices implementing RESTful API endpoints, service-to-service communication, and database integration layers that supported 10,000+ daily transactions across multiple product lines.
• Integrated Scikit-learn, XGBoost, and pandas-based machine learning models into production web applications for predictive analytics, building model serving endpoints and automated retraining pipelines that improved forecast accuracy to 85%.
• Designed and optimized Python-based ETL automation workflows orchestrated with Airflow and Apache Spark that processed and transformed 2M+ enterprise records daily, ensuring data integrity across downstream analytics dashboards and ML feature stores.
• Refactored and modernized 50+ legacy Java batch processing jobs, implementing error-handling frameworks and retry logic that reduced pipeline failure rates by 45% and improved system reliability for business-critical nightly data loads.
• Partnered with DevOps teams to implement CI/CD pipelines with Docker containerization, GitHub Actions, and Jenkins for automated testing, build, and deployment workflows, reducing release cycles from bi-weekly to daily.
• Led the migration of 30+ legacy Java-based services to Python, redesigning monolithic schedulers into modular scripts with improved logging, unit testing, and error handling, reducing average job execution time by 25%.
Education
The University of Texas at Arlington