Virtusa is hiring for Gen AI Developer at Chennai/Hyderabad location.
Pls find below the JD for your reference:-
Experience: 17-25 Years
Job Description
CLOUD AI SERVICES PRACTICE • MICROSOFT AZURE
Cloud AI Engineer Agentic AI & Generative AI (Azure)
We are looking for a senior engineer with 10 years of overall software/ML engineering experience, including at least 3 years building Agentic AI and Generative AI solutions on Microsoft Azure, to design, build, and deploy production-grade agentic and GenAI systems across the full AI stack.
Responsibilities
Architect and build GenAI solutions using Microsoft Foundry and Azure OpenAI, selecting models via the Foundry model catalog (OpenAI, Llama, Mistral, MAI)
Design and own Retrieval-Augmented Generation (RAG) pipelines using Azure OpenAI Embeddings, Azure AI Search, and Foundry IQ
Lead development of production AI agents using the Microsoft Agent Framework on Foundry Agent Service, with tool-calling via the Model Context Protocol (MCP) and multi-agent coordination via the Agent2Agent (A2A)
protocolDefine and enforce Azure AI Content Safety guardrails and lead Foundry evaluations before production release
Own MLOps automation with Azure ML / MLflow and drive fine-tuning of models on proprietary data
Mentor engineers and guide architecture decisions for agentic AI adoption across teams
Required Skills
- 10-15 years overall software/ML engineering experience, including at least 3 years hands-on with Agentic AI and Generative AI specifically
Deep hands-on experience with Microsoft Foundry and/or Azure OpenAI in a production environment
Robust grasp of embeddings, vector search, and Retrieval-Augmented Generation (RAG) architecture
Practical experience with the Microsoft Agent Framework, MCP tool-calling, and A2A multi-agent patterns
Strong Python/.NET and Azure fundamentals (Blob Storage, Entra ID, Container Apps)
Preferred
Microsoft Certified: Azure AI Fundamentals (AI-901) or Azure AI Apps and Agents Developer Associate (AI-103)
Experience with MLOps, CI/CD, and production model monitoring at scale