Software Engineer at CodeGen International

Architecting and delivering production-grade agentic AI infrastructure for the Lumos platform, an enterprise-grade SaaS solution powering intelligent software engineering workflows. Led end-to-end development of the AI Fullstack Application Builder — a full-stack, AI-powered application that takes a user requirement and autonomously builds complete full-stack applications end-to-end. Available as a desktop application (Electron + Python backend) and a web version (Next.js frontend). Orchestrates specialized multi-agent workflows covering requirements analysis, architecture design, UI/UX planning, frontend and backend code generation, database integration, and automated deployment — delivering a production-ready full-stack application from a single natural language prompt. Led end-to-end development of the Lumos Agent Management Dashboard — a full-featured SaaS platform enabling AI agent creation, session monitoring, MCP server connectivity, RAG knowledge base management, and a testing playground; praised in multiple client demos including a David Group Company presentation. Designed and built an AI-driven UI/UX design platform — enables users to describe product requirements in natural language and receive multi-page screen designs, complete UI/UX product plans, and AI-generated prototypes, with real-time sharing, commenting, and feedback for collaborative design review. Developed an agent-human collaboration workspace for the software development lifecycle (SDLC) — enabling specialized agents, manager agents, and human team members to collaborate on a shared workspace in real time, including agent-to-agent and agent-to-human communication, shared context/memory, task assignment, and structured work handoffs. Designed and built multi-agent orchestration pipelines using LangChain, LangGraph, and CrewAI, coordinating agents, tools, and data sources for automated software engineering tasks. Developed and integrated RAG-based knowledge bases with dynamic creation capabilities and external application connectivity, enabling scalable domain-specific LLM reasoning. Fine-tuned LLMs and optimized prompts for task-specific adaptation; managed memory, context handling, and vector database design to improve reasoning consistency. Conducted knowledge-transfer (KT) sessions for engineering and Tech Lead forums, covering agentic architecture design, AI agent development workflows, MCP server connectivity, and RAG system implementation — accelerating team-wide AI capability adoption. Built agentic AI chat applications with Python and Node.js backends; deployed and managed agents using PM2 and Microsoft Azure infrastructure.

Undergraduate Trainee Software Engineer at CodeGen International

Developed and maintained the Travelbox enterprise travel-tech platform for multinational clients, delivering full-stack modules and critical client releases in a high-pressure environment. Owned and delivered the complete Tour Module from inception — built full Angular front-end and Java Spring Boot back-end, migrating legacy systems to a modern, scalable web architecture. Contributed to Reservation Manager, Supply Manager, and multiple enterprise modules across the Travelbox platform, ensuring stability and feature parity across client environments. Played a key role in the Flight Centre go-live in Q1 2025 — implemented timezone-aware product search across DP, Flight, Hotel, and Transfer modules and enhanced transfer search location criteria dependency. Delivered major client releases, resolved critical production issues, and implemented feature enhancements across modules under tight deadlines and multinational client expectations. Mentored newly onboarded junior developers, accelerating their ramp-up on codebases, workflows, and product architecture. Maintained secure, scalable, and maintainable codebases adhering to Angular and Spring Framework best practices, with CI/CD pipelines managed via GitLab.