Priyanka MN

Intelligent Chatbot Development for Trex
• Designed and implemented a Proof of Concept (POC) using the Azure AI Portal to develop an intelligent chatbot for Trex, the world's largest manufacturer of wood-alternative decking products.
• Integrated Azure Cognitive Services to enable the chatbot to understand and respond to customer queries based on brochures, warranty cards, and other product documentation.
• Enabled customers to check order status and warranty claim updates through a conversational interface, improving self-service capabilities and reducing support load.
• Demonstrated the feasibility of leveraging Azure's AI offerings to streamline customer interactions and enhance user experience.

• Engineered an AI-driven platform to parse and interpret complex financial business rules from unstructured text.
• Utilized Google’s Gemini LLM for advanced NLP to extract entities, relationships, and intent.
• Modeled extracted data into a Neo4j Knowledge Graph for structured and connected representation.
• Implemented a Retrieval-Augmented Generation (RAG) pipeline to enable natural language querying of financial rules.
• Developed a full-stack application using Python (Fast API) to automate analysis and graph-based querying.

• Leveraged Langfuse to build a robust AI life cycle management system, optimizing the entire flow from model fine-tuning to prompt life cycle management
• Utilized Langfuse's advanced capabilities to track and manage prompt changes, ensuring seamless updates and real-time notifications(custom APIs built for this) to users whenever prompts are altered
• Implementing fine-tuning workflows that enhance model accuracy and adaptability, all while ensuring transparency and consistent performance
• This solution is to significantly improve user engagement and model responsiveness, driving smarter, data-driven decisions

• Utilizing a broad range of data science tools and techniques, I extract insights from complex medical data sets(Discharge Summary) and develop innovative solutions to drive business success
• AI Full Stack Development (Back-End),Data science, analytics, machine learning, deep learning , Azure Open AI
• Implemented Retrieval-Augmented Generation (RAG) concepts to build and optimize AI models, enhancing performance by effectively combining pre-trained models with real-time data retrieval for more accurate and contextually relevant outputs
• Used embedding models like BERT(to train models for each document type) Azure OpenAI’s GPT-4o and mini to extract the relevant information from the medical documents
• Additionally took up work on building Airflow DAGs and Plugins for the orchestration layer and database activities
• Integrated RAI principles into development of models focusing on accuracy , transparency and fairness

• Developing an Auto ML product for complete Machine Learning life cycle management
• Leveraged Azure AutoML to automate the model selection, training, and hyperparameter tuning process, significantly accelerating model development and deployment
• Enabling the creation of high-performance models with minimal manual intervention, enhancing efficiency and ensuring optimal results for predictive analytics and AI-driven solutions

• Working with Dell client for MTrack project
• M track is a web application to track the hardware materials and its life cycle management

• SIWEB is a ONE STOP portal where project details with respect to the Revenue, Cost and Efforts can be tracked from end to end

• Customer Centricity tool is an application which captures the feedback and enhancements to be done for the client companies for different projects
• Also, the Benefits of both client and customer is captured