✍️ Projects Portfolio
✍️ Project Philosophy
"Every project is an opportunity to learn, figure out problems, invent, and reinvent."
🚀 Self-Hosted GPU Inference Platform
Tech Stack:vLLM | Kubernetes (MicroK8s) | FastAPI | Qdrant | Docker | Prometheus | Grafana | Ollama
Objective: A self-hosted, GPU-accelerated AI platform deployed entirely from versioned Kubernetes manifests on a single NVIDIA GPU node. Serves an OpenAI-compatible LLM endpoint via vLLM, alongside Open WebUI and a FastAPI retrieval-augmented generation service backed by Qdrant vector search.
The ingest pipeline chunks and embeds text, PDF, DOCX and image uploads, falling back to OCR via Tesseract when a PDF's text layer is missing or too sparse, with page caps that fail fast instead of exhausting host memory. Ships a full observability stack (Prometheus, Grafana, Loki, Tempo, OpenTelemetry) and a CPU-only Docker Compose path using Ollama, so the whole system also runs on a laptop with no GPU. Deployment is driven by 20 numbered, idempotent shell phases plus a runbook: a clean Ubuntu host reaches a working platform in one command.
🔗 Distributed URL Shortener
Tech Stack:Python | FastAPI | PostgreSQL | Redis
Objective: A high-throughput redirection service with custom aliases, click analytics, authentication and rate limiting, holding sub-100ms latency under load testing. Designed the schema, indexing and cache strategy for high-volume concurrent reads, then profiled and removed the query bottlenecks that appeared under contention.
🌐 Campus Connect
🔗 Live
💻 Code Repository
Tech Stack:MERN Stack (MongoDB, Express.js, React, Node.js)
Objective:
Developed a Campus Connect web portal using MERN stack. It includes user authentication, admission form submissions, and admin management. Improved user experience with React and ensured efficient data management using MongoDB.
🌾 AgroSmart
💻 App Code
Tech Stack:Django | Python | Machine Learning | HTML | CSS | JavaScript
Objective:
Consolidated soil nutrient analysis, rainfall data, crop recommendations, and yield predictions using machine learning. The app helps farmers make informed decisions on crop selection and yield forecasts.
Features:
- Soil analysis
- Rainfall insights
- Crop recommendations
- Virtual Market for organic products
- Secure logout
🧘♂️ The Yoga Instructor
Tech Stack:Numpy | Matplotlib | OpenCV
Objective:
Implemented a pre-trained deep learning model to estimate body poses in real-time and predict yoga asanas. The system helps guide users in performing correct yoga poses by detecting and measuring angles.
🤖 LLAMA_Chatbot (Finderr)
💻 App Code
Tech Stack:Python | GPT 3.5 | JavaScript | HTML | CSS | Llama-Index | Django
Objective:
Finderr employs the LlamaIndex and RAG (Retrieval Augmented Generation) to enrich its responses with custom data sources. The chatbot assists users in finding college-related information, improving user interaction with swift, accurate responses.
✍️ Blogs Site
💻 App Code
Tech Stack:Python | Django | Bootstrap | SQLite
Objective:
A web application that allows users to create, publish, and manage blog posts. It offers a user-friendly interface for both administrators and visitors to interact with the blog content.
🎮 Games Info App
💻 App Code
Tech Stack:Flutter | Dart | API | Firebase
Objective:
API integration for authentication and UI design. This application provides a platform for users to browse PC games and read reviews.
Keep Exploring More Projects ✨