Headline

SE @ TCS | GenAI & ML Developer | Agentic AI Systems | Python | MCP | FastAPI | React

About

I’m an AI Engineer currently building production-grade GenAI systems at TCS. My work focuses on designing AI-driven root cause analysis platforms for large-scale codebases, combining retrieval-augmented generation (RAG), agentic workflows, and structured evaluation mechanisms. I specialize in building scalable FastAPI-based backends and asynchronous services that power complex LLM pipelines. From stateful agents with LangGraph to tool-integrated reasoning workflows, I focus on turning cutting-edge AI concepts into reliable, enterprise-ready systems. My goal is simple: bridge strong backend engineering with modern AI orchestration to build scalable systems that deliver measurable impact.

Location

Hyderabad

AI stack

langchain

langgraph

mcp

openai

Work experience

GenAI & Python Developer at Tata Consultancy Services, Feb 2025 - Present · 1 yr 3 mos, Hyderabad, Telangana, India, • Developed an AI-driven Root Cause Analysis (RCA) platform for the OpenAirInterface 5G codebase to fix code issues based on error logs, reducing manual debugging by 35%. Engineered a Contextual-RAG (CRAG) pipeline using FAISS and semantic embeddings to ground LLM outputs. Improved precision by 22% by implementing a custom reranking logic and a validation layer using semantic similarity to filter hallucinations in real-time. • Developed a multi-agent 5G Code Generation system using LangGraph and MCP,achieving near-total automation. Engineered a Graph-RAG engine using FAISS and Knowledge Graphs for iterative IE expansion across ETSI specifications. Integrated a self-learning session flow to interactively resolve ambiguities and automated repo-level validation to ensure memory safety and protocol compliance. • Implemented an agentic Wi-Fi test automation framework using state-machine-based orchestration to track coverage (RTM) using SQLite for efficient knowledge base storage. Boosted functional test coverage by 28% by utilizing autonomous agents to synthesize and execute missing edge-case test scenarios. • Fine-tuned a T5-small (60M) Encoder-Decoder Transformer for x86 to ARM assembly translation using HuggingFace and PyTorch. Implemented full-parameter fine-tuning with AdamW and teacher forcing to map instruction sets. Evaluated model performance using BLEU scores to ensure structural and token-level translation accuracy. • Engineered asynchronous REST APIs using FastAPI to expose stateful LangGraph pipelines and multi-agent workflows across all internal GenAI projects.

Summer Internship at Indian Railways, Jun 2023 - Jul 2023 · 2 mos, Prayagraj, Uttar Pradesh, India,  At the Signal and Telecom Department of Indian Railways, I had the privilege of participating in an intensive training program designed to expand participants' knowledge and expertise in various facets of data communication, networking, computer networks, and signaling technologies applied within the railway sector.  It not only broadened my technical expertise but also instilled a deep appreciation for the critical role of telecommunications and computer networks in enhancing the safety and efficiency of railway operations.

Education

Bachelor of Technology - BTech, Electronics and Communication Engineering, Madan Mohan Malaviya University of Technology

Public links

github: https://github.com/arshitshukla