Overview
Design, develop, deploy and operate AI agents, copilots and intelligent applications on Microsoft Azure AI Foundry, integrating them with enterprise data sources and supporting production-grade Generative AI use cases across Orbia business groups.
What you'll do
- Contribute to Azure AI Foundry reference architectures, reusable design patterns and engineering standards for agents, copilots and RAG-based applications.
- Evaluate models from the Foundry model catalog (Azure OpenAI GPT-4o, o-series, embeddings, open-source models) for accuracy, latency, cost, safety and business fit.
- Recommend agent orchestration patterns (single-agent, multi-agent, tool calling, human-in-the-loop) and prompt engineering standards.
- Align with the AI Architect on platform selection across Azure AI Foundry, Databricks Agent Bricks, Copilot Studio and Power Platform for each use case.
- Contribute to ARB submissions, technical decision records and architecture exceptions for Foundry-based solutions.
- Build, test and deploy AI agents using Foundry Agent Service, Prompt Flow, Semantic Kernel, LangChain or AutoGen.
- Implement Retrieval-Augmented Generation (RAG) patterns using Azure AI Search, ADLS Gen2, embeddings, hybrid search, chunking strategies, citations and re-ranking.
- Integrate Foundry agents with Microsoft 365, Microsoft Teams, SharePoint, Power Platform, ServiceNow, SAP and other enterprise systems through APIs, APIM and managed identities.
- Use Databricks to build and maintain upstream data preparation pipelines, vectorization and feature engineering using PySpark, Delta Lake, Unity Catalog and Vector Search.
- Implement CI/CD pipelines using Azure DevOps or GitHub Actions and manage environment promotion across DEV / QA / PROD with proper RBAC and approvals.
- Deploy supporting services to Azure App Service, Container Apps, AKS or Functions, following infrastructure-as-code patterns (Bicep/Terraform) where applicable.
- Collaborate with security and integration teams to design secure connectivity, secrets management (Key Vault) and identity (Entra ID, managed identities, service principals).
What you'll need
- 3–7 years of overall IT experience with at least 2+ years in Generative AI / LLM application development.
- Hands-on experience with Microsoft Azure AI Foundry / Azure AI Studio, including Foundry Agent Service, Prompt Flow, model catalog, project setup, deployments, evaluations and tracing.
- Strong experience with Azure OpenAI (GPT-4o, o-series, embeddings), Azure AI Search, Azure AI Content Safety and related Azure AI services.
- Strong Python development skills with experience in agent / GenAI frameworks such as Semantic Kernel, LangChain, AutoGen or LlamaIndex.
- Solid experience implementing RAG patterns: chunking strategies, embeddings, vector indexes, hybrid search, re-ranking, citations and groundedness evaluation.
- Secondary skill – Databricks: hands-on experience with PySpark, Delta Lake, Unity Catalog basics, Databricks Workflows, Vector Search, and awareness of Agent Bricks / Genie / AI-BI.
- Ability to build and maintain data preparation, embedding and feature pipelines on Databricks that feed Foundry-based AI applications.
- Working knowledge of the Azure ecosystem: ADLS Gen2, Key Vault, Entra ID, managed identities, API Management, Application Insights and Log Analytics.
- CI/CD experience with Azure DevOps or GitHub Actions.
- Good understanding of security and governance controls: RBAC, least privilege, audit logging, private endpoints, DLP, Microsoft Purview, data classification and SOX / ITGC.
- Exposure to responsible AI principles, prompt engineering, hallucination control, content safety, human approval gates and production release criteria.
- Experience preparing solution design documents, ARB-ready materials, runbooks and technical SOPs.
- Awareness of cost management for AI workloads: token usage, model cost, search cost, model endpoints and FinOps guardrails.
- Experience working in an international environment at a global company with cross-functional teams and business stakeholders.
- Excellent communication skills in English, both verbally and in writing.
- Self-starter, result-oriented, with strong problem-solving and analytical skills, able to work independently and collaborate with senior architects and business partners.
Nice to have
- Familiarity with Microsoft 365 Copilot extensibility, Copilot Studio and Teams app integration is a strong plus.
- Infrastructure-as-Code experience using Bicep or Terraform is preferred.
Details
- Location: Pune, India.
- Willingness to travel, if required, as per project and business needs.
Read the full description and apply on the company’s own careers page.