Introduced and taught students to Python programming with an AI-driven approach, delivering hands-on coding sessions to build practical skills, problem-solving abilities, and readiness for real-world AI applications.

This project explores the application of LLMs, prompt engineering, and Retrieval-Augmented Generation (RAG) to enhance fault detection and classification in electrical grids. By integrating vector databases for efficient context retrieval and evaluating multiple AI models (open-source and proprietary), the work also addresses AI security risks such as prompt injection and model hallucinations. The research aims to improve precision, recall, and F1-score while ensuring the reliability and safety of grid operations.

• Designed and implemented an AI system and chatbot with RAG integration using AWS S3, Lambda, OpenSearch, and API Gateway, providing agents with customized feedback based on client compliance knowledge articles and datasets aimed at automating of Quality Assurance processes for call center operations, reducing review time by 40%.

• Integrated Whisper for speech-to-text transcription of customer calls and deployed OpenAI TTS for voice-based response systems.

• Built and containerized the application code using Docker for seamless deployment.

• Developed an automated pipeline to trigger AWS Lambda on S3 events for processing data and knowledge articles.

• Designed end-to-end machine learning pipelines, including embedding generation, similarity matching, and semantic search, for analyzing customer-agent interactions.

• Expose APIs via AWS API Gateway for post-method interaction.

• Automated quality assurance workflows and designed voice analytics systems using Prompt Engineering to analyze customer-agent recordings and transcriptions, enhancing response accuracy by 90% and agent performance.