Athreyan Mohana Krishnan Sangeetha

• Deployed Purdue GenAI Studio, a service providing researchers with privacy-preserving access to open-source LLM models. • Refactored an open-source LLM UI using Svelte, Typescript, and Python to enhance the frontend user experience. • Enhanced GenAI Studio functionality by developing a user-specific RAG based on PostgreSQL deployed via Kubernetes. • Developed improved user authentication to enable scaling the service to the Purdue research community of over 300.

Software Developer worked on Cloud Technologies. •Used AWS EMR running Apache Spark for Data Engineering trade records. •Designed and built a Horizontally Scaling solution using AWS Lambda and EventBridge for settling trades done within the Fidelity ecosystem. The Business team was able to get faster reports with improved system reliability. •Built several subsystems using AWS Lambda, Amazon SNS and SQS that supported error reporting activities. •Prototyped a calculation engine in Python whose updated version now exists in the current billing solution.

Natural Language Processing • Used Python regex and other libraries like VADER, spaCy to perform Natural Language Processing on documents to retrieve relevant financial data. Work done laid a basis for the future of the project. • Trained an ML model to identify important Financial data with a 95% accuracy and 99% specificity.

Web Development • Used ReactJS to build a reporting tool that took data and presented a tailored report based on user input. • Built an internal inventory tool using HTML, JavaScript, PHP, MySQL and MongoDB. The tool allowed GUVI to manage and maintain their digital learning catalogue.