My Projects

Developed a comprehensive dental treatment platform for Zsmile's innovative teeth alignment device. The solution includes a patient-facing mobile app that uses AI-powered video scanning to assess treatment eligibility, an IoT integration for device management and treatment monitoring, and a web-based CRM for orthodontists and technicians. Built the mobile application using Expo React Native with video processing capabilities that extract frames for ML analysis. Created native plugins in Swift and Kotlin for face detection functionality due to lack of compatible cross-platform libraries, and maintained these plugins throughout the project lifecycle. Developed Java Spring Boot backend services and a separate Express/Node.js backend for the CRM system. Collaborated with embedded systems team to ensure seamless app-device communication. Managed the complete AWS cloud infrastructure hosting the Java backend, Python ML services, and CRM components.

Developed a medical diagnostic platform for healthcare professionals to detect respiratory diseases using AI-driven video analysis. The application captures 30-second patient recordings and employs pose detection algorithms to analyze breathing patterns frame by frame, identifying potential respiratory conditions. Built the mobile application in Expo React Native with integrated pose detection capabilities supporting both vertical and horizontal video recording. Created native plugins in Swift and Kotlin for pose detection functionality due to lack of compatible cross-platform libraries, and maintained these plugins throughout the project lifecycle. Developed the Node.js backend server and implemented database architecture using Supabase to store analysis results and AWS S3 references for video storage. The complete server infrastructure is hosted on Azure.

Developed an advanced background search platform that performs comprehensive investigations on individuals using name, phone, email, or keywords. The system aggregates data from multiple sources, including social media APIs, web scrapers, public records databases, and criminal record checks. Implemented intelligent data processing using LLM technology via LangChain to normalize, summarize, and synthesize information from disparate sources into cohesive, actionable reports. Built the web application using Next.js with Supabase for user management and report storage per user. Integrated multiple third-party APIs and search services, including the Google Search API, for finding relevant articles and web presence.

Developed an intelligent security automation platform designed for on-premises infrastructure that revolutionizes incident response by reducing manual analysis time for security analysts. The system connects to Wazuh (open-source SOC manager) via webhooks to receive security alerts and employs AI agents to autonomously investigate and respond based on pre-defined decision graphs. Built the core orchestration service that processes incoming security alerts and manages automated response workflows. Work also on The agent manager component (built in Golang) to coordinate autonomous security responses. Implemented human-in-the-loop checkpoints before critical actions to ensure safe autonomous operations. Developed the Next.js dashboard with TypeScript featuring real-time alert monitoring, critical alert notifications, visual graph charts showing action workflows, a graph builder for creating custom response scenarios, and comprehensive metrics from Wazuh agents. Integrated MCP connections to Slack and Gmail to automatically notify administrators and security managers of critical alerts requiring human intervention, ensuring timely response to high-priority security incidents. Designed and deployed the complete infrastructure using Docker containerization, Kubernetes orchestration, and CI/CD pipelines for automated deployments.

Developed a full-stack RAG (Retrieval-Augmented Generation) chat system that enables intelligent conversations about user-provided articles using Claude AI. The system features a hybrid microservices architecture combining Go and Node.js for optimal performance. Built a Go-based API gateway with PostgreSQL for user authentication and Redis for response caching. Implemented a Node.js RAG service using LangChain.js, FAISS vector store for local embeddings, and Claude AI for natural language generation. Created a React 19 frontend with Shadcn/UI components and Tailwind CSS v4 for a modern ChatGPT-style interface. Key features include article processing via URL extraction, text chunking and embedding generation, semantic search with vector storage, persistent conversation history, and intelligent caching for repeated queries. The system supports 10-50 concurrent users and achieves sub-100ms response times for cached queries. Designed complete Docker containerization with multi-stage builds and Docker Compose orchestration for seamless deployment.

Designed and implemented a complete DevOps pipeline for a Next.js face detection application using industry best practices. Built a custom Next.js application with real-time webcam face detection using face-api models to detect landmarks, age, gender, and emotions. Architected AWS infrastructure using Terraform including VPC, subnets, security groups, and EC2 instances for Kubernetes master and worker nodes. Deployed separate production (3 pod replicas) and staging (2 pod replicas) environments with proper network isolation and security configurations. Implemented comprehensive CI/CD pipelines using GitHub Actions with three main workflows: automated submodule updates, quality checks (TypeScript, ESLint, formatting), and continuous deployment. Configured Docker multi-stage builds with automated pushes to Docker Hub, triggering webhook-based Kubernetes rolling updates with zero downtime. Set up Kubernetes cluster with NGINX Ingress Controller, SSL/TLS certificates, ConfigMaps, and health checks. Automated deployment updates using kubectl with SSH key authentication.

A comprehensive e-commerce solution that demonstrates modern DevOps practices and full-stack development capabilities. The application features a complete product management system with admin and user roles, shopping cart functionality, and real-time inventory tracking. The project showcases CI/CD implementation with Jenkins pipelines that automatically build, test, and deploy the application. GitOps principles are implemented using ArgoCD for Kubernetes deployments, ensuring declarative infrastructure management and automated synchronization. The tech stack leverages Next.js 15 with React 19 for a modern, performant frontend, PostgreSQL for robust data management, and a complete containerized infrastructure using Docker and Kubernetes. Monitoring is handled by Prometheus for production-grade observability.

This project implements a Flask API server that integrates PostgreSQL database operations with the Model Context Protocol (MCP), providing a unified interface for both traditional CRUD operations and AI-powered natural language database queries. The server leverages LangChain and Ollama to enable users to interact with their database using plain English queries, which are then translated and executed as SQL operations. The application demonstrates modern backend development best practices with a clean, modular architecture separating concerns across configuration, middleware, services, and routes. Security is prioritized through JWT-based authentication, rate limiting to prevent abuse, input validation using Marshmallow schemas, and comprehensive security middleware. The project includes protection against common vulnerabilities and implements proper error handling throughout. Testing excellence is a cornerstone of this project, featuring 72+ tests across multiple categories including unit tests, integration tests, end-to-end tests, security tests, stress tests, regression tests, and sanity checks. The test suite achieves high code coverage and uses modern testing tools like pytest-xdist for parallel execution. The entire stack is containerized with Docker, making deployment and development environment setup straightforward and consistent.

UrbanHive is a comprehensive mobile social platform designed to address critical challenges faced by modern residential communities. Originally conceived to solve the problem of organizing night watch operations in areas with security concerns, the project evolved into a full-featured community engagement application. The application tackles three main problems: the absence of organized systems for managing night watch operations, difficulties faced by new residents in connecting with local communities, and the limitations of traditional social media platforms for community-specific needs. UrbanHive provides a dedicated, purpose-built solution with location-based filtering and community-specific features. Built as a capstone project by a two-person team, UrbanHive demonstrates sophisticated technical implementation, including a React Native mobile client supporting both iOS and Android platforms, a Python Flask RESTful API backend, and MongoDB for flexible data management. The architecture follows MVC patterns with a service layer design, ensuring scalability and maintainability.

Developed a modern, responsive landing page for S.O.D.A (Software Outsourcing Development Architecture), a software outsourcing company specializing in AI and cybersecurity solutions for startups and entrepreneurs in the medical and cybersecurity sectors. Built the website using Next.js with React and TypeScript, featuring comprehensive sections including company project portfolio, partner showcase, technical expertise overview, team member profiles, and company values. Implemented responsive design ensuring optimal viewing experience across all devices. Deployed on Vercel for reliable hosting with automatic CI/CD integration. The website serves as the primary digital presence for the company, effectively communicating S.O.D.A's capabilities and successful client projects to potential customers.