Dhanya Sri Vasantha

AI Engineer with hands-on experience in designing and deploying Generative AI and Agentic AI solutions using Large Language Models (LLMs), prompt engineering, Model Context Protocol (MCP), and Retrieval-Augmented Generation (RAG) frameworks. Proven expertise in building intelligent AI agents capable of reasoning, planning, and interacting with external tools using LangChain and LangGraph, while implementing multi-step workflows such as ReAct for complex task execution. Strong experience developing semantic search and knowledge retrieval systems using vector databases including FAISS, with optimized embedding strategies for accurate, grounded responses.

Proficient in Python, SQL, Azure, and AWS, with additional strengths in machine learning, data engineering, and scalable cloud deployments. Adept at delivering secure, production-ready AI systems that improve automation, decision-making, and operational efficiency across enterprise environments.

AI Engineer at Siemens (2025-07 – Present)

• Designed and deployed enterprise-grade agentic AI systems using LangGraph and GPT-4.1 to modernize fragmented business workflows, resulting in 40% faster response performance and improved reasoning accuracy across internal applications.
• Built scalable RAG pipelines integrated with vector databases, APIs, and enterprise knowledge sources to solve low-context response issues, improving answer relevance and reducing latency by 30% in production environments.
• Developed Azure Data Factory pipelines aligned with Medallion Architecture to process high-volume enterprise data, handling 500K+ records daily while achieving 99% data reliability through SCD Type 2 and Data Vault models.
• Engineered an AI-powered document intelligence platform using LLMs and NLP to automate extraction from complex business documents, reducing manual effort by 60% and accelerating turnaround time for downstream teams.
• Implemented resilient AI operations frameworks with automated retries, exception handling, monitoring, and dead-letter queues to stabilize production pipelines, reducing processing failures by 45% and improving uptime.

Data Scientist at Trion IT Solutions (2021-10 – 2023-04)

• Led the development of enterprise AI solutions across NLP, recommendation engines, and predictive analytics to improve customer experience, increasing user engagement and model adoption across multiple business functions.
• Built end-to-end MLOps frameworks using MLflow, DVC, and CI/CD pipelines to address inconsistent model releases, reducing deployment cycles by 40% and improving reproducibility by 35%.
• Developed large-scale data processing pipelines using Pandas and PySpark for model training and analytics workloads, reducing processing time by 45% and improving compute efficiency by 30%.
• Designed transformer-based and deep learning models for fraud detection, forecasting, and content generation use cases, delivering production-ready AI systems that improved prediction quality and automation outcomes.
• Implemented model monitoring solutions with Power BI dashboards to provide real-time visibility into accuracy, drift, and business KPIs, enabling proactive decisions and improving stakeholder reporting efficiency by 50%.

Master's in Data Analytics Engineering – George Mason University

Bachelor's in Computer Science and Engineering – Devineni Venkata Ramana & Dr. Hima Sekhar MIC College of Technology