Om Balgude — Full-Stack Engineer specializing in Real-Time Systems and AI Products


I build scalable, AI powered products from the ground up. As a Full Stack Engineer, I specialize in TypeScript, Node.js, and Python, engineering everything from real time WebSockets to complex RAG workflows.
Extreme ownership and high velocity are my core principles. I leverage AI tools to accelerate development cycles, ensuring that every feature I ship is fast, secure, and built to scale.
Languages
Frontend
Backend
Tooling
Backend AI Engineer Intern | @FlyRankAI
- •Engineered AI orchestration pipelines for a production-grade CLI platform supporting 5+ intelligent commands, integrating 2 LLM providers to automate developer workflows across file management, system operations, and content generation.
- •Designed and developed 10+ RESTful API endpoints and backend services using Node.js, Express.js, and TypeScript to power modular AI workflows.
- •Streamlined orchestration logic and request routing across 4+ backend workflows, strengthening response consistency, maintainability, and overall system reliability.
- •Accelerated feature delivery by embedding AI coding agents into the development loop, cutting implementation and debugging time while preserving production-grade code quality.
Technologies Used
Full Stack AI Engineer Intern | @Sumago Infotech Pvt. Ltd.
- •Delivered an AI-powered Standard Operating Procedure (SOP) Assistant by integrating OpenAI, Google Gemini, and Retrieval-Augmented Generation (RAG) for context-aware enterprise knowledge retrieval.
- •Developed and maintained 12+ RESTful API endpoints and backend services using Python and FastAPI, supporting AI integrations through a modular backend architecture.
- •Owned frontend development for the SOP Assistant and Krishi AI using Next.js, shipping 20+ reusable UI components wired into backend services.
- •Tuned RAG retrieval strategies and inference pipelines to improve response quality and performance, adopting AI-assisted workflows to raise development velocity.
Technologies Used
Full Stack Developer Intern | @AI Vidya
- •Built the AI Vidya platform end-to-end, pairing a responsive React frontend with scalable Node.js/Express backend services.
- •Developed 10+ RESTful API endpoints using Node.js, Express.js, and MongoDB, enabling secure, efficient frontend-backend communication.
- •Applied object-oriented design principles to structure modular, reusable components and maintainable application architecture, while improving performance through optimized API interactions and database query tuning.
Technologies Used
React Native mobile app with a comprehensive fitness engine and AI generation.
An AI-powered fitness copilot delivered as a React Native mobile app, backed by a TypeScript/Express API and a Next.js site, unified through a Turborepo.
- •Built ARC, an AI-powered fitness copilot delivered as a React Native mobile app, backed by a TypeScript/Express.js API and complemented by a Next.js site.
- •Engineered 25+ RESTful API endpoints on PostgreSQL and Drizzle ORM across a 17-table schema, powering workout planning, nutrition tracking, and habit management.
- •Unified the ecosystem through a Turborepo monorepo with 7 shared packages for end-to-end type safety.
- •Developed a reusable fitness engine with a 40+ exercise library for personalized, AI-driven workout generation and macro calculations shared across the app and backend.
Technologies Used
Secure browser-based code execution platform powered by isolated Docker sandboxing.
A full-stack online coding playground enabling real-time execution of 5 programming languages inside isolated Docker containers with secure sandboxing.
- •Built a browser-based Remote Code Execution platform supporting 5 programming languages (Python, Rust, Go, JavaScript, TypeScript) through an interactive, Monaco-powered web IDE.
- •Leveraged AI coding agents to rapidly prototype and iterate on the execution pipeline and container orchestration logic.
- •Applied rigorous manual review to the security-critical sandboxing components to ensure safe execution.
- •Secured untrusted code execution with strict resource caps (100 MB memory, 0.5 CPU, 5-second timeout) and full network isolation to prevent abuse.
Technologies Used
AI-assisted certificate verification platform with OCR analysis and blockchain trust.
A 6-service microservices platform combining OCR pipelines, automated verification workflows, and optional blockchain anchoring.
- •Architected a 6-service microservices platform (React, Express.js, FastAPI, MongoDB, Redis, Nginx) for automated certificate verification.
- •Built a document analysis pipeline with OpenCV and Tesseract OCR to extract 10+ certificate fields, cutting estimated manual review effort by 70% to 80%.
- •Anchored certificates on-chain with Solidity while designing a graceful fallback to database-based verification when blockchain services were unavailable.
Technologies Used
Real-time collaborative drawing platform with a Turborepo monorepo architecture.
A real-time collaborative drawing platform supporting multi-user editing with synchronized WebSocket events and a persistent database backend.
- •Architected a real-time collaborative drawing platform using a Turborepo monorepo, integrating 2 applications with 6 shared packages to enable scalable full-stack development, code reuse, and end-to-end type safety.
- •Engineered 6 RESTful API endpoints and 6 WebSocket event types using Node.js, TypeScript, PostgreSQL, and Prisma ORM, enabling real-time collaboration, room management, authentication, and low-latency canvas synchronization.
- •Built an interactive collaborative canvas supporting multiple drawing tools with synchronized multi-user editing through an event-driven WebSocket architecture, delivering a responsive real-time collaboration experience.
- •Designed a secure authentication system with JWT-based access and refresh tokens, httpOnly cookies, and bcrypt password hashing to protect REST APIs and persistent WebSocket connections.
- •Developed a relational PostgreSQL database using Prisma ORM to manage users, collaborative rooms, and persistent drawing history, enabling seamless state restoration for newly connected participants.
Technologies Used
Full-stack P2P payment platform with ACID-compliant transaction workflows.
A secure peer-to-peer payment platform built with React and Node.js featuring atomic debit and credit operations through MongoDB sessions.
- •Engineered a full-stack P2P payment platform using React, Node.js, Express.js, and MongoDB, implementing secure authentication, account management, and peer-to-peer fund transfers.
- •Designed 7 RESTful API endpoints secured with JWT authentication, bcrypt password hashing, and Zod runtime validation, ensuring protected financial transactions and robust request validation.
- •Built a responsive React application with 13 UI components (8 reusable components across 5 application views), delivering a consistent user experience for authentication, account management, and money transfers.
- •Implemented ACID-compliant transaction workflows using MongoDB sessions to execute atomic debit and credit operations, guaranteeing data consistency and preventing race conditions during concurrent fund transfers.
- •Designed a MongoDB data model with 2 collections and 1:1 ObjectId relationships, optimizing user-account mapping while enforcing unique indexing for efficient authentication and account lookups.
Technologies Used
Open to full-time opportunities, freelance projects, and collaborations focused on scalable products, AI-powered applications, and modern web experiences.
