Snigdha Mohana Addepalli

• Optimized MongoDB collections with strategic indexing to improve query response times and reduced LLM hallucinations by implementing current date-time parameters in prompts for accurate recent statistics analysis. Conducted comprehensive testing of file transfer query functionality across four distinct scenarios, identifying critical issues with file name-specific queries and documenting expected system behavior for proper data visualization and reporting. • Developed conversational AI chatbot using LangGraph Agent framework with ChatGPT integration for weather and stock market information. Implemented enterprise-grade AI practices including chain-of-thought reasoning, structured outputs, and safety guardrails for production deployment. • Designed comprehensive service layer endpoints for automated AI agent generation platform with PostgreSQL database architecture and time-series partitioning strategies for high-volume metrics storage. Built real-time monitoring systems tracking network traffic, latency, and success/failure rates across MCP servers and AI agents, creating analytics endpoints with data visualization capabilities for production observability and performance optimization.

• Engineered and optimized traffic sign detection and road quality assessment models using Python, PyTorch, and TensorFlow, achieving 80% accuracy and reducing false positives by 30%. Improved model detection accuracy by 15% over baseline models and cut data retrieval time by 50% with efficient AWS S3 storage. • Collaborated with cross-functional teams to implement and deploy AI-driven solutions for traffic management and infrastructure assessment, enhancing model impact on real-world applications.

• Developed a machine learning model for accurate health condition diagnosis based on patient symptoms, achieving 100% accuracy with the Random Forest classifier and 97.6% with Gradient Boosting and Decision Tree. Leveraged Joblib for efficient model deployment. • Improved user experience by designing an interactive GUI with Gradio, allowing for seamless interaction and reliable condition diagnosis.

To-Do List Webpage Used Web Development and Data Sciences to create the TO DO LIST app which can help set the tasks and give reminders for any planned task a user has. It also provides an option to remove the task once it has been completed.

Coursework: CS 5010 - Programming Design Paradigm CS 5100 - Foundations of Artificial Intelligence CS 5800 - Algorithms CS 6120 - Natural Language Processing CS 6140 - Machine Learning CS 7180 - Advanced Perception CS 5330 - Pattern Recognition and Computer Vision CS 6620 - Fundamentals Cloud Computing