Portfolio
A collection of projects spanning AI/ML, full-stack development, data engineering, and open-source contributions. Built with purpose, shipped to production.
Disaster Response AI Copilot
Disaster Response AI Copilot
Agentic AI copilot for real-time emergency response. Built on Azure OpenAI (GPT-4o) and Azure AI Foundry with a Retrieval-Augmented Generation pipeline backed by Azure Cognitive Search. Streamlit UI provides situational-awareness dashboards and actionable decision support during natural disasters. Python data pipeline in data_pipeline.py feeds live feeds into the RAG index.
Crypto Trade Prediction
Crypto Trade Prediction
Automated cryptocurrency trading system built with seven Docker Compose microservices: backfill (historical OHLCV data ingestion), realtime (live WebSocket trades), candle_maker (aggregates ticks into candles), predict (ML price prediction), trade_bot (automated order execution), train (scheduled model retraining), and streamlit_app (live visualization dashboard). Cronjob scheduling coordinates the pipeline. Deployed on DigitalOcean.
Heart Disease Prediction with Generative AI
Heart Disease Prediction with Generative AI
Medical data analysis platform combining an ML classification pipeline (app.py + data_pipeline.py + eda.py) with an OpenAI-powered natural-language explanation layer. Patients receive a risk score plus a plain-English breakdown of contributing clinical factors. Fully containerised with Docker Compose and deployed live on Streamlit Cloud.
Car Price Prediction — MLflow + Kafka + Debezium
Car Price Prediction — MLflow + Kafka + Debezium
Complete MLOps system for predicting used-car prices (Ford dataset). Architecture: Streamlit UI → FastAPI backend → PostgreSQL → Debezium CDC → Kafka → ML Service → MLflow (experiment tracking + model registry) → MinIO (artifact storage). Docker Compose orchestrates all services including Zookeeper, Kafka UI, Adminer, and MinIO Console. scikit-learn models served via MLflow with Plotly charts in the Streamlit frontend.
Movie Recommendation — Cosine Similarity & Collaborative Filtering
Movie Recommendation — Cosine Similarity & Collaborative Filtering
Movie recommendation engine deployed on AWS EC2. FastAPI backend serves cosine-similarity recommendations computed from a PostgreSQL-backed dataset. Streamlit frontend (Dockerfile.streamlit) renders results with Pandas/Plotly. Docker Compose manages the backend, Streamlit app, and database services. Pre-processed data loaded via sql_load.py.
NYC Taxi Ride Time-Series Prediction
NYC Taxi Ride Time-Series Prediction
Batch ML pipeline using Hopsworks feature store, DagsHub + MLflow for tracking, time-series models, and a live Streamlit deployment for ride demand forecasting.
Data Science Salary Prediction Platform
Data Science Salary Prediction Platform
End-to-end MLOps pipeline: Apache Kafka streaming, Airflow orchestration, PySpark processing, Snowflake warehouse, Redis caching, PostgreSQL, and React frontend with Flask API. Deployed on AWS EC2.
Heart Disease Prediction System
Heart Disease Prediction System
Production ML system with MLflow experiment tracking, DagsHub integration, Streamlit UI deployed on DigitalOcean with Docker. High-accuracy ensemble classification.
Taxi Data Prediction Dashboard — Snowflake + Snowpark
Taxi Data Prediction Dashboard — Snowflake + Snowpark
Taxi demand analytics platform using Snowflake as both the data warehouse and compute engine. Snowpark Python UDFs run ML transformations directly in Snowflake. Jupyter notebooks (notebooks/) handle EDA; SQL scripts (sqls/) define transformations; Python src/ manages the pipeline. Streamlit Cloud hosts the interactive dashboard.
Customer Segmentation & Sales Analysis
Customer Segmentation & Sales Analysis
Interactive R Shiny application (app.R) for statistical learning and business intelligence. Applies unsupervised ML for customer segmentation, visualises sales trends, and drives product recommendations. Persistent data in SQLite3; rsconnect directory enables deployment to shinyapps.io. Live demo at vaibhavbansal.shinyapps.io/sdmproject.
Natural Disaster Prediction (VIT Capstone)
Natural Disaster Prediction (VIT Capstone)
VIT University capstone (4 GitHub stars): multi-hazard prediction system covering earthquakes (Japan 2016, Nepal 2015), floods (Chennai 2015, Kerala 2018, Hoppers Crossing), hurricanes (Atlantic/Pacific), tsunamis (Japan 2011, Indonesia 2018), and cyclones. scikit-learn (Random Forest) + TensorFlow (earthquake_model.ckpt) models served via Flask. Google Maps API visualises predicted impact zones. YouTube demo linked.
Realtime Logs Processing — Airflow + Kafka + Elasticsearch
Realtime Logs Processing — Airflow + Kafka + Elasticsearch
Production-grade streaming log pipeline: Confluent Kafka ingests application events, AWS MWAA (Managed Apache Airflow) orchestrates the DAGs, and Elasticsearch indexes logs for full-text search and analytics. AWS Secrets Manager handles all credentials securely. Topics tagged: python, aws, elasticsearch, airflow, kafka, aws-secrets-manager, mwaa.
Medallion Architecture — Azure Data Engineering
Medallion Architecture — Azure Data Engineering
Enterprise lakehouse on Azure following the Bronze/Silver/Gold medallion pattern. Bronze layer ingests raw data via Azure Data Factory pipelines into Azure Data Lake Gen 2 (Parquet format). Silver layer applies DBT transformations in Azure Databricks. Gold layer writes refined aggregates to Azure SQL Database. Azure Key Vault manages secrets throughout. Topics: azure, dbt, databricks, azuredatafactory.
NYC Taxi Data Engineering & Power BI Visualization
NYC Taxi Data Engineering & Power BI Visualization
Serverless AWS data engineering pipeline for NYC taxi analytics. AWS Lambda processes raw trip data; EventBridge triggers scheduled runs; AWS Glue Studio performs ETL with a visual data catalog; Athena queries the S3 data lake. Power BI and AWS QuickSight deliver stakeholder dashboards. Topics: python, aws-lambda, powerbi, aws-athena, aws-glue, event-bridge.
DevSecOps Movie Booking — Terraform + Jenkins + EKS
DevSecOps Movie Booking — Terraform + Jenkins + EKS
Full DevSecOps pipeline for a React movie-booking app. Stage 1: Terraform provisions EC2 and Amazon EKS (eksctl, t3.medium nodes). Stage 2: Jenkins CI/CD (20+ plugins) runs SonarQube SAST quality gates → OWASP dependency check → Trivy container image scanning → Docker build/push → ArgoCD GitOps deploy to Kubernetes. Prometheus + Grafana + Node Exporter stack handles observability. 2 GitHub stars.
DevSecOps Food App — Full CI/CD Pipeline
DevSecOps Food App — Full CI/CD Pipeline
DevSecOps pipeline for a Swiggy food-delivery clone (React frontend). Jenkins pipeline: GitHub webhook → SonarQube code quality gate → OWASP dependency check → Trivy image scan → Docker build/push (vaibhavbansal26/swiggy:latest) → container deploy on port 3000. Terraform provisions the AWS EC2 instance. Prometheus + Grafana + Helm for monitoring.
DevSecOps App — EKS + Prometheus + Grafana
DevSecOps App — EKS + Prometheus + Grafana
DevSecOps pipeline for a BookMyShow clone (React + CSS). Jenkins CI/CD: GitHub → SonarQube SAST → Docker Scout vulnerability scan → Docker build/push → ArgoCD GitOps → Amazon EKS. AWS CloudFormation provisions infrastructure. Helm manages Kubernetes releases. Prometheus + Grafana observability stack with Node Exporter. 1 community fork.
Music Streaming — 3-Tier MERN on AWS EKS
Music Streaming — 3-Tier MERN on AWS EKS
3-tier microservice music streaming platform. Frontend: React/Vite + Tailwind CSS + Context API. Backend: three TypeScript/Node.js/Express services (userService, adminService, songService) backed by MongoDB and NeonDB (PostgreSQL). Redis caching, Cloudinary for audio/image assets, JWT authentication. Docker Compose for local dev; Jenkins CI/CD pipeline (SonarQube, OWASP, Trivy) deploys to AWS EKS via Kubernetes manifests.
Amazon Clone with Stripe Payments
Amazon Clone with Stripe Payments
Full-stack Amazon clone built with Next.js and React (JavaScript). Firebase handles authentication and Firestore data. Stripe integration processes payments. Redux manages global cart state. Tailwind CSS styles the UI. Deployed on Vercel at amazon-clone-stripe-ten.vercel.app. 1 GitHub star.
MERN Instagram — Photo Sharing App
MERN Instagram — Photo Sharing App
Full-stack Instagram clone with MERN architecture. Node.js/Express backend (app.js) uses MVC pattern with separate models, routes, middleware, and config directories. React client handles the feed, stories, follow/unfollow, post likes, and comments. MongoDB persists user profiles and posts. JWT-based auth middleware. JavaScript 91%, CSS 7%.
MERN YouTube Application
MERN YouTube Application
Full-stack YouTube clone using the MERN stack. Monorepo structure with client (React SPA) and server (Node.js/Express REST API) directories. Supports video upload to the uploads/ directory, video playback, user authentication, channel subscriptions, likes, and comments. MongoDB persists all data. JavaScript 95%, CSS 3%.
Funnel / Website Builder — DashClicks
Funnel / Website Builder — DashClicks
Production drag-and-drop funnel and website builder built for DashClicks India Ltd. Powered by React and GrapesJS, enabling marketers to build landing pages without code.
Video Conferencing App
Video Conferencing App
Real-time video conferencing app built with Node.js, Socket.IO, and EJS server-side rendering. Users create or join rooms via unique IDs; the server (server.js) brokers WebRTC peer connections. EJS templates (views/) render the UI; public/ serves client-side JS and CSS. Deployed live on Render.
React Native Real Estate App
React Native Real Estate App
Cross-platform real estate browsing app built with React Native (Expo) and TypeScript. Appwrite provides backend services (auth, database). NativeWind + Tailwind CSS handles styling. Features property listings, search filters, and detailed property screens. TypeScript 97%, built with nativewind-env and metro config. Topics: typescript, react-native, authentication, tailwindcss, appwrite, nativewind.
Pizza Delivery & Realtime Order Tracking
Pizza Delivery & Realtime Order Tracking
Real-time pizza ordering system with Node.js, EJS templating, and Socket.IO for live order-status push updates. MVC structure: app/ (controllers/models), routes/, resources/, public/. Webpack mix bundles SCSS assets. Procfile enables Heroku deployment. JavaScript 57%, EJS 34%, SCSS 9%.
Natours — Travel Booking App with Stripe
Natours — Travel Booking App with Stripe
RESTful travel booking API and web app built with Node.js, Express, MongoDB, and Mongoose. MVC architecture: controllers/, models/, routes/, utils/, views/ (Pug templates). Features JWT + cookie-based auth, tour CRUD, user reviews, Stripe payments, email notifications, and interactive map. ESLint + Prettier configured. 1 GitHub star.
GrapesJS Advance Components (npm)
GrapesJS Advance Components (npm)
Published npm package (grapesjs-advance-components) that extends GrapesJS with 15 production-ready drag-and-drop blocks: FAQ, Headings (h1–h6), Paragraph, List, Link, Image, Video, Progress Bar, all HTML Input types (text, email, checkbox, radio, file, etc.), Dropdown, Textarea, Map, Divider, Button, and Slider. Available via CDN or npm install. 5 GitHub stars, 1 fork.
VSCode Extension — Search Blogs
VSCode Extension — Search Blogs
VS Code extension (vb-search-blog) that surfaces the latest tech blogs directly inside the editor sidebar. Built with the VS Code Extension API in JavaScript. Includes a .vscodeignore, CHANGELOG, and structured test suite. Lets developers stay in flow without switching to a browser.
28 projects · 95+ repos on GitHub →