My chatbot-style personal portfolio powered by a RAG model

Inspiration

I first built a static personal portfolio, but it felt dull and unengaging—even to me as the creator. If I wasn’t finding it interesting, recruiters or visitors wouldn’t either. While exploring ways to make it better, I considered adding a chatbot. That’s when I came across some amazing interactive personal portfolios. Inspired by them, I decided to create something different: a chatbot-style portfolio powered by a RAG application. Unlike typical themed portfolios, mine responds intelligently to a visitor’s natural language queries. Combined with interactive elements, it transforms a portfolio into an engaging, memorable experience.

What it does

This project is an AI-powered personal portfolio that lets visitors interact with my resume, projects, and background through natural conversation. Instead of browsing static pages, users can ask questions like “What projects have you worked on?” or “What skills do you have in AI/ML?” and get accurate, conversational answers. The chatbot retrieves relevant information from my resume and structured JSON data, making it a dynamic, user-driven experience.

How we built it

The frontend was built and deployed with Vercel. I manually integrated the RAG application using NVIDIA Build’s free nvidia-nemotron-nano-9b-v2 LLM API and Pinecone as the vector database. To give it a professional touch, I used my GitHub Student Pack’s free domain offer through Namecheap and hosted it at akashbellary.me. And i used Google nano banana image generating model to generate my avatar and other elements in the portfolio.

Challenges we ran into

The biggest challenge was integrating multiple APIs and ensuring smooth communication between the chatbot, LLM, and vector database. While not overly difficult, it required careful debugging and optimization to achieve a seamless experience.

Accomplishments that we're proud of

• Successfully built a unique AI-powered portfolio that stands apart from static websites.
  
• Integrated a working RAG pipeline that retrieves precise, resume-based answers in real time.
  
• Deployed the project on a custom domain (akashbellary.me), making it production-ready and professional.
  
• Created an engaging experience that not only showcases my profile but also demonstrates my AI/ML engineering skills.
• I deployed this on Thursday [25th sept] and have used this to apply to the companies and it already helped me get call from Dave AI company recruiter.
  

What we learned

• How to design and integrate a RAG workflow into a practical application.
  
• Best practices for combining frontend deployment (Vercel) with backend AI services.
  
• The importance of user experience in tech projects—making it functional is one thing, but making it engaging is what really leaves an impact.
  
• Leveraging free resources like NVIDIA APIs and GitHub Student Pack perks effectively.
  

What's next for My Chatbot-Style Personal Portfolio powered by a RAG model

• Enhance the RAG system to handle queries more dynamically and contextually.
  
• Integrate my GitHub repositories, so visitors can directly ask questions about my projects (e.g., “How do I run CampaignIO locally?”).
  
• Add a direct contact form where users can send me a message tagged with their name, delivered straight to my email.
  
• Implement a calendar scheduling feature, allowing freelance clients or recruiters to directly book a meeting based on my availability.
  

Built With

• github
• google-nano-banana
• javascript
• namecheap
• nvidia-nemotron-nano-9b-v2-llm-api
• pinecone
• rag
• react
• tailwindcss
• vercel