Hi, my name is Nandhi Kanti Vinay Kumar
I'm a Full Stack Software Developer. I do what I mentioned in my Resume

"Chat with knowledge base"
Prep Pal is a sophisticated Retrieval-Augmented Generation (RAG) application using cutting-edge technologies such as LangChain, Python, FAISS(Facebook AI similarity search), GPT-4, and MongoDB. This project is designed to provide an end-to-end solution for job seekers. This interview coaching chatbot pairs up with a resume generation tool that I previously worked on. Users can upload their resumes, job descriptions, and URLs of career sites containing interview questions and answers. The chatbot then contextualizes this information to generate the most relevant and accurate answers to the interview questions and provides responses based on user's resume (i.e. work experiences, projects) helping users prepare effectively for their interviews.

The system utilizes a multi-stage RAG (Retrieval-Augmented Generation) architecture. All unstructured data is stored as chunks in a vector database. Information is retrieved from the vector database using vector similarity search based on the user's query. Addtionally, "MapReduce" loads and summarizes the context, ensuring that the chatbot delivers quick and accurate responses by performing asynchronous API requests to the language model. This method minimizes irrelevant context and significantly improves chatbot's performance and reliability, reducing the API costs.

Pulse Events is an event hosting and management platform designed to streamline the process of organizing and attending events. This project showcases a comprehensive integration of modern technologies such as Next.js with TypeScript, Spring Boot, MongoDB, and Google Cloud Platform, and best practices in web development. Proactively resolved challenges in optimizing server performance and securing user data and overall application.

The backend uses Spring Boot for a robust and scalable foundation, with secure JWT authentication and optimized MongoDB queries for enhanced performance. Well-developed RESTful APIs contribute to faster response times.
The frontend is built with Next.js, featuring reusable React components for a dynamic and responsive interface. Server-side rendering and effective caching improve page load speeds, ensuring a smooth user experience. The design is mobile-responsive for accessibility on any device.
In DevOps, Docker and GitHub Actions enable continuous integration and deployment (CI/CD), facilitating quick and reliable deployments. The application is deployed to a Kubernetes cluster on Google Cloud Platform (GCP) for scalability and high availability.

MyCarrerCoach is a comprehensive Generative AI resume buider tool built for university students. It takes minimal input from the users to build the resume, analyzes resumes for factors like keywords, formatting, and content, providing feedback and suggestions for improvement. It aims to increase the chances of a resume getting noticed by employers and ultimately securing interviews.

Collaborated with a team of developers and Computer Science department at SFSU, identifying students' resume challenges. Developed methods and Flask APIs in Python, integrated GPT to generate high-quality resumes. The resume achieved consistent scores of 70+/100 on screening tools. Implemented a NLP based benchmarking and scoring system to evaluate each resume sections and provide feeback and suggestions, resulting in improvement of content quality. Formulated and optimized PostgreSQL queries and designed complex JSON data storage and indexing for flexible resume structures. Delivered comprehensive user experience (UX) by implementing 100% reusable React components, custom hooks, and Context API, enhancing code reusability and maintainability. Proactively communicated progress in bi-weekly SCRUM meetings with the team and product manager, showcasing initiative in finding solutions.

"Picture Perfect" is a image sharing and digital marketplace platform with combining elements of Imgur but with additional features such as image editing and marketplace features. Users can explores various categories of images including space, wildlife, nature, film, animals, and more. Addtionally, users can upload images to the specific categories and can crop, re-size, edit, add filters before uploading it. They can also buy/sell images from other users.

Built with React.js, Django, MySQL, TailwindCSS, MaterialDesign, Figma. Utilized containerized docker images, built CI/CD pipelines, facilitating automatic deployments to Amazon Web Services, leveraged Amazon RDS for streamlined database management. Collaborated with a team of 5, followed end-to-end Agile SDLC, lead weekly SCRUM meetings and provided updates on the product status regularly. Demonstrated Agile project management and development skills.