Inspiration
As a technologist and hackathon builder, I’ve always wanted my portfolio to speak for me — not just about me. Most portfolios are static, but I envisioned one that thinks, talks, and adapts. That’s how FarhunVerse was born — an AI-powered interactive portfolio that merges my personality, technical expertise, and creativity into one immersive experience.
FarhunVerse isn’t just a website — it’s my digital twin, powered by AI and designed to reflect my journey from hackathon builder to professional engineer at HCL. It represents my core belief:
“Technology isn’t about replacing people — it’s about augmenting potential.”
What it does
FarhunVerse is an AI-powered personal portfolio and interactive storytelling platform built entirely with Streamlit, featuring intelligent integrations that make it far more than a static resume.
Here’s what it does:
💬 AI Resume Navigator (FarhunBot): An LLM-powered chatbot built with LangChain + OpenAI that allows users to chat with my resume. It answers any question about my skills, projects, and achievements — as if I were explaining it in person.
🧠 Tech Showcase: A dynamic dashboard that presents my hackathon-winning projects (FusePay, NEARVision, GuardianAI, EduRegion Explorer, Cryptographic Farming, etc.) with category filters, Lottie animations, live demos, GitHub links, and video showcases.
🌐 Beyond the Code: A personal touch — a section that visually narrates my journey outside technology — from football, hackathons, and modelling to music and fitness — each with glowing animated gradient cards and auto-playing slideshows.
⚙️ Live Analytics & Engagement: A real-time visitor counter, feedback system, and AI-driven presentation flow to make the portfolio self-aware and continuously improving.
📩 Integrated Email System: Built-in contact form with SMTP, allowing direct communication with me from within the site.
In short —
💫 FarhunVerse is an AI-enhanced fusion of personality, portfolio, and performance.
How I built it
FarhunVerse was built with a modern AI + frontend + data stack focused on simplicity, responsiveness, and intelligence.
Tech Stack:
🧱 Frontend: Streamlit (custom CSS, animations, interactive layouts)
🧠 AI Layer: LangChain + LangChain-Community + LangChain-OpenAI
🧩 Vector Store: FAISS for semantic resume embedding
🔍 Embeddings: text-embedding-3-large from OpenAI
⚡ LLM Backend: GPT-4o-mini integrated via ChatOpenAI
🗂️ Data Processing: PyPDF2 for text extraction from resume
🎨 Animations: Lottie + CSS glow & motion
📡 Email System: Python smtplib + email-validator
☁️ Deployment: Streamlit Cloud (runtime pinned to Python 3.10 for dependency stability)
Each page (Home, Tech Showcase, AI Resume Navigator, Beyond the Code, and Contact) was crafted for purposeful storytelling, interactive engagement, and emotional connection — combining data visualization, AI conversation, and modern design.
Challenges I ran into
Every innovation starts with a few bumps — and FarhunVerse was no exception.
⚠️ Dependency Conflicts in LangChain: Multiple breaking changes across langchain-core, langchain-community, and langchain-openai required fine-tuned version control and API migration.
🧩 OpenAI Embedding Error (proxies issue): Faced client initialization conflicts in the OpenAI API due to newer SDK architecture. Resolved by restructuring the embedding logic and decoupling proxy arguments.
🎨 Streamlit Layout Limitations: Streamlit’s rigid layout system required creative use of CSS, HTML, and markdown for fluid animations and glowing borders.
💾 Performance Optimization: Cached Lottie animations and vector embeddings to reduce cold-start time on Streamlit Cloud.
☁️ Deployment Runtime Issues: Encountered FAISS installation issues under Python 3.13 — fixed by forcing runtime to 3.10 and rebuilding environment manually.
Each challenge made the final build more robust, modular, and production-ready.
Accomplishments that I am proud of
💬 Created an AI chatbot that understands and explains my resume contextually.
🌟 Built a fully responsive, animated portfolio using only Streamlit + CSS — no frontend frameworks.
🧠 Integrated LangChain + FAISS + GPT-4o-mini locally and deployed successfully on Streamlit Cloud.
🪩 Designed a unique Beyond the Code section — connecting personality to profession.
📊 Combined data visualization, storytelling, and AI into a single portfolio experience.
⚙️ Solved complex dependency and API-level issues that even production teams struggle with.
Most importantly —
🏅 FarhunVerse is now my AI-powered professional identity, live and evolving.
What I learned
How to build AI-integrated web apps using LangChain, FAISS, and Streamlit seamlessly.
Real-world handling of OpenAI SDK updates and breaking changes.
How design and interactivity can make a technical portfolio feel human.
Best practices for environment isolation, dependency pinning, and model version control.
The importance of combining technical precision with creative storytelling in hackathon projects.
What's next for FarhunVerse - Where Code Meets Creativity
The journey doesn’t stop here. 🚀
Here’s what’s coming next:
🤖 Multi-Agent Intelligence: Integrate prebuilt LangGraph AI agents for real-time code, project, and resume reasoning.
🎙️ Voice Interaction: Add voice-based Q&A for conversational resume exploration.
🧩 AI-Powered Portfolio Generator: Let others build their own AI portfolios based on this framework.
💡 Analytics Dashboard: Track user engagement and sentiment for adaptive personalization.
🌐 Blockchain Badge Verification: Integrate decentralized credentials for project and certification proof.
Ultimately, FarhunVerse will evolve into a fully interactive AI-driven ecosystem — where anyone can experience the synergy of creativity, intelligence, and innovation in one click.
Built With
- faiss
- faiss-local-vector-index
- langchain
- openai
- pypdf2
- smtplib
- streamlit
- streamlit-lottie
- tiktoken
- watchdog
Log in or sign up for Devpost to join the conversation.