YouHired is an AI-powered developer portfolio and resume platform built to solve a real problem: developers rebuild the same information — projects, skills, experience — separately for every portfolio, resume, and job application. YouHired centralizes it into one profile and generates everything else from that single source. The platform is built as two coordinated applications — a Next.js marketing site and a full MERN-stack dashboard app — sharing one backend for data, transformations, and generation. How it works: A user connects their GitHub and fills out a structured profile once. From that single data source, YouHired generates a live portfolio, an ATS-optimized resume, and an AI-written bio grounded in the user's actual project history — not generic filler text. A recommendation engine then surfaces weekly job matches based on the user's real skills and preferences, and everything — portfolio and resume alike — can be exported or shared with one link. What I built: A centralized profile schema that transforms one data source into multiple structured outputs (portfolio + resume) A resume-generation pipeline that normalizes skills and formatting into ATS-friendly output An AI bio generator that pulls from the user's real projects/skills instead of generic prompting A weekly job-recommendation system matched to user profile data A template-driven rendering layer so layouts can change without touching underlying data GitHub API integration to auto-sync and display real project history

Finoryx started as a personal need: spreadsheets weren't cutting it for tracking recurring bills, shared expenses with friends, and actual spending patterns over time. I built it as a full-stack personal finance platform that handles all three in one place. The core is a dashboard that tracks monthly budgets, account balances, and spending against a set budget target in real time. Recurring transactions — daily, weekday-only, monthly, or yearly — are automatically processed on schedule rather than requiring manual entry every time, and a group expense feature lets shared costs (rent splits, group purchases, utilities) be tracked and settled the way Splitwise handles shared expenses. What I built: A recurring-transaction engine supporting custom schedules (daily, Mon–Fri, monthly, yearly) with automatic daily processing An expense analytics layer with category breakdowns, spending trend charts, and period-over-period comparisons A group expense system for splitting and settling shared costs among multiple people A transactions table with search, filtering, and CSV export/import A quick-add expense flow supporting natural-language input (e.g. typing "800 petrol from HDFC yesterday" and having it parsed automatically)

Khiladi Adda is a real-money gaming (RMG) platform supporting multiple games — Ludo, Rummy, Ludo Tournament, Cricket Tournament — with the full compliance and financial infrastructure real-money apps require: KYC onboarding, wallet management, and region-aware legal compliance. I worked across the platform's backend and its Rummy-specific product line (RummyAdda), plus a standalone sibling app (Bharat Rummy) built entirely backend-first. Across these, I handled everything from onboarding and wallets to performance optimization at scale. User-facing systems: KYC onboarding — Aadhaar/PAN verification workflows for secure user registration Wallet system — real-time balance updates, deposits, withdrawals, and winnings transfers, with full transaction logs Location-based compliance — auto-detects a user's state/country via IP to enforce region-specific gaming rules TDS compliance — automated tax deduction handling on winnings, per Indian regulations Multiple payment gateway integrations for deposits and withdrawals Performance and infrastructure: Batch API processing cut latency by 40%; Redis caching brought response times down to ~4-5ms Replaced manual Excel exports with a Google Sheets streaming pipeline, supporting 100,000+ records and saving roughly ₹1 lakh in infrastructure costs Built and optimized REST APIs to handle high-traffic gaming loads without degradation Security: Password hashing (Bcrypt) and data encryption (Crypto) across all three applications Rate limiting on external integrations and sensitive endpoints JWT-based authentication across user sessions I also contributed to the React.js-based admin panel used for real-time platform monitoring.

This is a backend system for a full-scale Learning Management System, built to support real course delivery — not just CRUD operations. It handles three distinct user roles (admin, instructor, student), each with different permissions and access to course, payment, and progress data. The system covers the full lifecycle of an online course: enrollment and payment, lecture progress tracking, wishlisting, and certificate generation on completion — all behind a versioned, secured API layer. Core systems: Role-based access control (JWT + Bcrypt) with distinct permission sets for admin, instructor, and student roles Razorpay integration for secure course payments Cloudinary + Multer for video/image uploads (course content, thumbnails) Progress tracking across lectures, with wishlist management Automatic certificate generation on course completion Production practices: Centralized logging (Winston + Morgan) for activity and error tracking Security hardening: Helmet, HPP, rate limiting, and encrypted credentials Versioned API structure (/api/v1/) for clean long-term scalability

EventPulse is an anonymous event posting and application platform, built to handle real-time interaction — chat, notifications, and applications — at scale. I worked on this as part of the engineering team at [Company], focused primarily on the platform's real-time infrastructure and performance layer. The platform runs on a Node.js microservices architecture, distributing load across specialized services to support around 10,000 daily active users without single points of bottleneck. What I worked on: - Real-time data streaming — used Kafka to process live notifications and chat data, increasing overall system throughput by 70% - Live chat infrastructure — built with Socket.io for near-instant messaging with robust concurrency handling. - Caching layer — implemented Redis for frequently accessed data, cutting average response times by roughly 60%. - Notification system — integrated Firebase push notifications with a priority queue for time-sensitive alerts, achieving a 99.9% delivery rate Supporting systems (team-built): - JWT-based authentication with Bcrypt password hashing and Crypto encryption. - MySQL with Sequelize ORM for consistent, transactional data handling Razorpay integration for in-platform payments

EazeAccounts is a financial management platform built for a real client — a Chartered Accountant needing QuickBooks-equivalent functionality for managing US accounting workflows across multiple clients and organizations. The platform handles the core financial documents an accountant actually needs: balance sheets, income statements, and a full Chart of Accounts, alongside multi-bank account tracking and transaction management. Every transaction feeds into real-time balance updates and detailed financial reports, so users always have an accurate, current view of their finances. Core systems: - Bank integration via Plaid, supporting multiple linked accounts per user - Sales tracking, income and balance sheet management, and Chart of Accounts (COA) support - Customizable financial reporting - Role-based and organization-based authentication (JWT), supporting multi-organization access - Stripe integration for payment processing Architecture: RESTful API design for secure, efficient handling of sensitive financial data Sequelize ORM for optimized handling of complex financial queries at scale