You will be taken to Avanade’s own careers page to apply. Job ID R00356080.
Job description
Role Summary
Avanade India’s Data & AI practice is looking for an Agentic AI & Copilot Studio SME to lead client enablement and adoption for agent-based AI solutions built on Microsoft Copilot Studio and the wider Microsoft agent ecosystem. This is a training-led, client-facing role: the primary focus is designing and delivering structured skilling — instructor-led workshops, hands-on labs, train-the-trainer programmes and champion communities — for large organisations building and scaling agents.
What distinguishes this role from a platform-training role is agentic depth. You are expected to understand how agents actually work — orchestration and reasoning loops, tool and function calling, grounding and retrieval, memory and context handling, multi-agent patterns, evaluation and guardrails — and to teach Copilot Studio in that context rather than as a feature tour. You will build reference agents yourself, advise on live engagements, and be the person client teams escalate hard agent-design questions to.
Key Responsibilities
Own the design and delivery of agentic AI and Copilot Studio skilling programmes for large enterprise audiences — curriculum design, instructor-led sessions, hands-on labs, self-paced content, assessments and certification pathways.
Run train-the-trainer and champion-enablement programmes so client teams can sustain and scale agent development independently after the engagement.
Teach agent design as a discipline, not a tool walkthrough — task decomposition, orchestration and reasoning patterns, tool/action design, grounding strategy, memory and context, human-in-the-loop checkpoints, failure modes and cost/latency trade-offs.
Tailor content and depth to distinct audiences — business makers, pro-code engineering teams, and business and IT leadership.
Build reference and demo agents in Copilot Studio — topics, generative answers, actions and connectors, agent flows, knowledge sources and multi-agent handoffs — to anchor training and seed client use cases.
Design and maintain lab environments, sandbox tenants and exercise sets, keeping them current with the fast-moving agent-platform roadmap.
Provide hands-on advisory on live engagements: agent design reviews, grounding and knowledge-source guidance, action/connector patterns, prompt and instruction design, testing and evaluation approaches, and go-live readiness.
Advise on when to build in Copilot Studio versus a pro-code approach on Azure AI Foundry or an agent framework — and how the two coexist in a single client landscape.
Establish evaluation and Responsible AI practice for agents: test sets and regression suites, groundedness and quality measurement, content safety, guardrails, human oversight and monitoring in production.
Advise clients on agent governance and operating models — environment strategy, DLP policies, maker onboarding, publishing and approval workflows, agent lifecycle management, licensing considerations and usage analytics.
Support use-case discovery workshops — identifying, qualifying and prioritising agent scenarios, and setting realistic expectations on what agents can and cannot do reliably.
Act as the practice reference point for agentic AI enablement: build reusable training assets, playbooks, agent templates and accelerators, and mentor consultants joining the capability.
Contribute to presales as the subject-matter expert — enablement approach, effort and duration estimates, demos and client-facing solution walkthroughs.
. Required Skills & Experience
9–12 years of overall experience in software engineering, automation, data or AI roles, with a recent and substantial focus on generative and agentic AI.
Hands-on experience designing and building agentic AI solutions — multi-step orchestration, tool and function calling, retrieval and grounding, memory and context management, and multi-agent collaboration patterns.
Up to 2 years of hands-on experience with Microsoft Copilot Studio, covering both building agents and delivering training on the platform.
Practical Copilot Studio depth: topics and trigger design, generative answers and knowledge sources, actions and connectors, agent flows, publishing across channels (Teams, web, M365 Copilot), and testing and troubleshooting agents.
Demonstrated experience designing and delivering technical training to large organisations — instructor-led workshops, hands-on labs, train-the-trainer and champion programmes, for both business and technical audiences.
Working knowledge of the wider Microsoft AI stack: Azure AI Foundry / Azure OpenAI Service, Azure AI Search for retrieval, and how these complement low-code agents.
Familiarity with agent frameworks and protocols — Semantic Kernel, AutoGen, LangChain/LangGraph, MCP or similar — sufficient to position and compare approaches credibly.
Understanding of prompt and instruction design, RAG patterns, and agent evaluation — groundedness, task success, regression testing and quality measurement.
Understanding of Responsible AI in practice: content safety and filtering, guardrails, data boundaries, auditability and human oversight.
Working knowledge of the Power Platform (Power Automate, Dataverse) and of Microsoft 365 data sources as grounding and action surfaces for agents.
Comfort with code — Python or C#, plus REST APIs and integration patterns — enough to build custom actions, explain pro-code extensions and debug integrations.
Strong facilitation and communication skills, with the confidence to hold a room of senior stakeholders as well as a lab of hands-on makers.
Ability to produce enablement collateral independently — decks, lab guides, quick-reference material and recorded walkthroughs.
Preferred / Good to Have
Microsoft certifications: Azure AI Engineer Associate (AI-102), Power Platform Fundamentals (PL-900) or Developer (PL-400), or Copilot/agent-related Microsoft credentials.
Experience with declarative agents, the Microsoft 365 Agents Toolkit, Graph connectors or custom engine agents for deeper extensibility scenarios.
Experience taking agents to production — observability, cost and latency optimisation, versioning and lifecycle management.
Exposure to non-Microsoft agent platforms (AWS Bedrock Agents, Google Vertex AI Agents, open-source frameworks) for comparative positioning.
Experience across multiple industry domains (manufacturing, BFSI, consumer goods, energy, healthcare, or others).
Experience in a global delivery / SI environment (Avanade, Accenture, or similar) working with distributed teams across geographies.
Prior contribution to reusable training assets, agent accelerators or practice-building initiatives.
