Headline

GenAI Application Engineer | LLMs · LangGraph · Multi-agent orchestration · RAG · FastAPI | Building AI-powered products | AI Engineer @ TCS

About

I build Generative AI products that actually ship. As an AI Engineer at TCS, I design and develop production-grade LLM applications — multi-agent systems, RAG pipelines, vision-powered AI tools, and full-stack applications that go from idea to deployed product. My current focus: Generative AI, Agentic AI, LangGraph multi-agent architectures, RAG, Google ADK, and MCP — building practical systems that solve real problems, not just demos. Some things I've built: → House Inspection AI Assistant — LangGraph multi-agent pipeline with Llama 4 Scout vision model, FastAPI backend, React frontend, deployed on Render + Vercel → Meeting Transcription Chatbot — RAG system with Qdrant, HuggingFace embeddings, and Groq for natural language Q&A over meeting content I work across the full AI stack — prompt engineering, vector databases, backend APIs, cloud deployment, and frontend integration. Fast learner. Bias for building. Always experimenting with what's next in AI. 🔗 GitHub: github.com/suprajasribalaji

Location

Chennai

AI stack

claude code

langfuse

langgraph

lovable

Google ADK

Qdrant

Work experience

Engineer in AI at Tata Consultancy Services, Oct 2025 - Present · 10 mos, - Architected a House Inspection AI Assistant using LangGraph multi-agent pipeline (defect classifier + report writer nodes), integrating Groq's Llama 4 Scout vision model for image analysis and Llama 3.3 70B for report generation. - Built a Meeting Transcription Chatbot with end-to-end RAG pipeline using LangChain, Qdrant vector database, HuggingFace embeddings, and Groq API for low-latency LLM inference. - Migrated LLM backend from Google Gemini to Groq API across multiple services, implementing optimized rate limiting and async FastAPI endpoints. - Designed normalized PostgreSQL schemas with SQLAlchemy ORM, implementing JWT authentication, bcrypt password hashing, and real-time input validation. - Deployed full-stack AI applications on Render (backend) and Vercel (frontend) with Firebase storage integration and environment-based CORS configuration.

Assistant System Engineer - Trainee at Tata Consultancy Services, May 2025 - Present · 1 yr 3 mos, - Completed foundational training across Linux, Networking, and software engineering practices at TCS. - Began transitioning into Generative AI and agentic systems, self-initiating projects in LangChain, LangGraph, and RAG pipelines.

Foundations of AI at Edunet Foundation, Apr 2025 - May 2025 · 2 mos, Hybrid

Intern - Full stack developer at 5C Network, Oct 2023 - Apr 2024 · 7 mos, Coimbatore, Tamil Nadu, India, Frontend Development • Improved performance and Maximized UX by fixing bugs and redesigning UI using ReactJS with ViteJS, and React Native. • Streamlined desktop app navigation by implementing routing with ElectronJS, improving interaction and access across three pages. • Enhanced user engagement and journey by displaying a welcome modal once per day and optimizing button flows with ReactJS and ViteJS. • Supported radiologists by adding subtitles modal for easier report management, significantly improving usability and workflow. • Expanded language accessibility using i18n in ReactJS, broadening the platform’s reach and enhancing user comfort and ease of use. Backend Development • Implemented RESTful APIs while migrating two microservices from ExpressJS to NestJS across three modules, enhancing functionality and managing dependencies effectively. • Optimized report generation timelines by redesigning message triggers and extending service hours using SNS, SQS, ExpressJS and MongoDB, ensuring timely email notifications and improved efficiency.

Education

Bachelor of Engineering - BE, Computer Science, Dr.N.G.P. Institute of Technology.

HSC, Rajan Matriculation Higher Secondary School

Public links

github: https://github.com/suprajasribalaji

leetcode: https://leetcode.com/suprajasri89