Vivek Nagar
GenAI Engineer | Agentic AI & LLM Platforms | AI Product Builder | Full-Stack Developer
Hyderabad
mail.viveknagar@gmail.com
I'm a GenAI and ML Engineer with 4+ years of experience building production AI systems across insurance and financial services. My work spans agentic AI, LLM orchestration, RAG, MCP, model fine-tuning, evaluation, and cloud-native deployment. At TCS, I've worked on enterprise AI platforms, including an agentic QA automation system that reduced a 2–3 day manual workflow to approximately 4 minutes, and a deep-learning PII/NPI masking platform. Beyond enterprise engineering, I build digital products, including RunRestro, a cloud-based restaurant POS and management platform with AI-powered menu digitization, and TCML, a medical education platform available on iOS and Android. I enjoy turning practical problems into scalable products using AI, Python, Django, Flutter, and cloud technologies.
Independent Software Developer & AI Product Builder
- • Build and develop independent digital products that solve practical problems in restaurant operations and medical education.
- • Develop RunRestro, a cloud-based restaurant POS and management platform featuring AI-assisted menu digitization, order and table management, QR-based ordering, billing, kitchen workflows and reporting.
- • Develop TCML, a medical learning platform for MBBS students and NEET PG/FMGE aspirants, with organized study resources, notes, video lectures and exam-preparation content.
- • Build full-stack applications using Django REST APIs, Flutter, PostgreSQL, cloud infrastructure, authentication and asynchronous background processing.
- • Work across product architecture, backend APIs, mobile applications, deployment, and ongoing feature development.
GenAI & ML Engineer
TCS
- • Designed and deployed enterprise agentic AI systems using LangGraph, LangChain, MCP, FAISS and RAG to automate QA workflows.
- • Reduced a manual QA workflow from 2–3 days per user story to approximately 4 minutes through automated test-script generation.
- • Built multi-stage retrieval pipelines with metadata-based tenant isolation, hybrid search and cross-encoder reranking.
- • Implemented LLM observability and tracing with LangSmith, plus confidence-gated regeneration and retrieval evaluation.
- • Worked on PRISM, a deep-learning PII/NPI detection and masking platform, achieving under 0.1% false negatives and a 62% improvement in name-masking effectiveness over legacy rule-based approaches.
- • Integrated cloud-based AI pipelines, asynchronous APIs and data processing using Azure services, Celery, Redis, PySpark and AWS infrastructure.
MCP-Powered Agentic QA Platform
Built an enterprise agentic AI platform for Zurich Insurance that transforms JIRA stories into generated Robot Framework test scripts. The solution combines LangGraph orchestration, MCP tools, FAISS retrieval and LLM-based generation. It reduced a 2–3 day manual workflow to approximately 4 minutes, with retrieval constraints and tracing to improve reliability.
RunRestro is a cloud-based restaurant POS and management platform designed to simplify restaurant operations. It combines AI-assisted menu digitization from menu photos, QR-based ordering, table and order management, billing, kitchen order tickets, thermal printing and business analytics. The goal is to reduce manual work and help restaurants manage their operations from mobile devices and standard POS hardware.
TCML (The Charsi of Medical Literature) is a medical education platform built for MBBS students and competitive exam aspirants preparing for NEET PG, FMGE and related examinations. The mobile application organizes learning resources such as notes, video lectures, question banks and exam-preparation content. I worked on the full-stack platform, including the Django REST backend, Flutter mobile applications, secure access controls and cloud deployment. The app is available on iOS and Android.
AI-Driven Claims Submission & Estimation
• Working on an AI-driven insurance claims submission and estimation solution that processes broker-submitted emails and associated insurance documents to extract relevant claim and submission details. • Developed an AI-powered document extraction workflow using field-specific prompts and predefined ground-truth values to improve the accuracy and consistency of structured data extraction from forms and supporting files. • Configured field-level extraction instructions and validation references to capture the required information from heterogeneous insurance documents and map it to the appropriate submission fields. • Integrated extracted information with the downstream submission application, enabling insurance details to be populated and used in the estimation workflow. • Contributing to the automation of manual insurance submission processing by converting unstructured broker communications and document content into structured data for downstream insurance estimation.
M Tech
NIT - Bhopal · A+
Claude Architech
Claude
Microsoft Azure AI Fundamentals (AI-900)
Microsoft
LoRA/QLoRA Fine-Tuning (LLM)
Udemy
Scikit-learn & ML Fundamentals
Udemy
Attendee · Oct 24, 2026