Skip to content
Search jobs inside your AI agent. Claude, ChatGPT, Claude Code, Codex, Gemini, Cursor and Grok
Connect

SCC Services Romania

NewBe an early applicant

Senior Agentic AI Engineer / Tech Lead

Posted

Summary

Senior engineer / tech lead in SCC's Agentic AI & Automations CoE in Bucharest, designing and building AI agents and automations for IT service-management transformation on Microsoft Azure, Microsoft Foundry and Copilot Studio. Owns architecture, engineering standards, integrations (ServiceNow, Microsoft Graph, Dynamics 365), IaC/DevOps pipelines and agent quality for customer solutions.

What you'll own

  • Build and improve the AI agents for our Service Management and Service Delivery transformation.
  • Work with the architects on the architecture and solution design for each customer, from the first options to the reference implementation the team builds from.
  • Try new releases and new agent technologies in short proofs of concept to keep our solutions up to date and innovative.
  • Act as the technical authority on how the CoE builds, alongside the architects: engineering standards, design patterns, reviews of designs and code, and the decision on how each feature gets built.
  • As the Senior Technical authority, Leads engineers on complex technical challenges; leads technical solutions and implementations. Leads technical discussions with customers.
  • Build the shared product core on Microsoft Foundry and Azure, and the automations behind it: the decision logic, prompts and tools every customer's agents share, exposed through the MCP, and automations that act through Microsoft Graph, Exchange Online, Intune, other infrastructure and business resources and CRMs, ERPs, including on customers' own servers.
  • Make our agents and automations work with the AI platforms customers already run, through MCP, the A2A protocol and REST APIs, wherever each platform supports them.
  • Build, support and improve the agents and automations already running for customers on our partner's Agentic AI platform: deterministic and agentic workflows, API-based integrations, on-premises execution agents, knowledge sources, and channels such as Microsoft Teams and the service portal.
  • Own agent quality: evaluation test sets that gate every release, regression runs before model changes, and tracing of every conversation and tool call.
  • Build and maintain the integrations with ServiceNow (Table and Service Catalog APIs, OAuth 2.0, domain separation), Microsoft Dynamics 365, and other business systems.
  • Own the engineering pipeline: Bicep templates and Azure DevOps pipelines with automated tests, Copilot Studio agents packaged as Power Platform solutions, and release rings across customer tenants through Azure Lighthouse.
  • Build security in: managed identities, Key Vault, least-privilege Graph and Exchange permissions, allow-lists of what each automation may change, and an audit trail for every action.
  • Monitor agents and automations, track model tokens per customer, and lead the technical response when something breaks.


What you'll bring

Required

  • At least six years designing and delivering solutions on Microsoft Azure, including at least three shaping AI or agent-based solutions that run in production.
  • Seniority and technical self-sufficiency: you take a problem from a rough outline to a working, supported solution on your own, without detailed technical guidance, and you know which decisions to bring to the architects.
  • AI solution design: you've designed and delivered AI agents or applications built on large language models on Microsoft Foundry or Azure OpenAI, and you can explain your choices on models, tool calling, retrieval, evaluation, security and cost.
  • Solution design for enterprise customers: turning business processes and constraints into designs, with the trade-offs on security, data residency and cost written down.
  • Deep knowledge of Azure's integration, automation and identity services: Functions, API Management, Service Bus or Event Hubs, Azure Automation, Key Vault and Application Insights, with Microsoft Entra ID workload identities and Microsoft Graph.
  • Hands-on delivery: infrastructure as code, deployment pipelines, and enough scripting and code to build and review the hard parts yourself.
  • Agent technology beyond Microsoft: MCP and A2A, and at least one other agent framework or SDK, such as LangChain and LangGraph, the OpenAI Agents SDK or Google's Agent Development Kit.
  • Hands-on experience with at least one commercial conversational or agentic AI platform besides Microsoft's, including low-code workflow and agent builders, API-based integrations, knowledge ingestion and channel deployment.
  • Technical leadership: you've set technical direction for other engineers, reviewed their designs and work.
  • Professional English at C1 level (advanced).


Nice to have

  • Copilot Studio and Power Platform, with application lifecycle management (ALM) through solutions and the Power Platform command-line interface (CLI).
  • Microsoft Agent Framework (Semantic Kernel or AutoGen experience counts), and building MCP servers.
  • Agent platforms on other clouds, or the AI agents customers run in ServiceNow, Google AI Enterprise or Salesforce Agentforce.
  • Evaluation and observability for AI: test sets, automated scoring of answers, and tracing with OpenTelemetry or similar tools.
  • Hybrid estates: Azure Arc, Hybrid Runbook Workers, on-premises Active Directory and Exchange hybrid.
  • ServiceNow integration: Scripted REST APIs, Integration Hub, domain separation.
  • Voice AI: Teams Phone, contact centre platforms or speech services.


Certifications

We look for current Microsoft certifications in both Azure and Azure AI: Azure AI Apps and Agents Developer together with at least one of Azure AI Cloud Developer, DevOps Engineer or Azure Solutions Architect. Expert-level AI credentials add weight, such as Multi-Agent AI Solutions Expert or Agentic AI Business Solutions Architect Expert, and Microsoft AI Agent Builder.


Who does well here

You're a senior professional and work autonomously: give you a goal and the constraints, and you'll shape the design, build and bring back a working result, without needing someone to steer each step. Engineers and architects seek you out because your reviews make their work better. You're at home in Azure and Microsoft's AI stack, and curious enough about the rest of the field to know when another tool does a job better and when it only looks newer. When a customer's environment breaks one of your assumptions, you fix the design first and the build second.


What success looks like

By the end of your first three months:

  • You've reviewed the current build and agreed new engineering standards with the architects and engineers.
  • With the architects, you've shaped the solution design for at least one customer, and a first part of the shared product core runs in our test environment, with evaluation test sets gating its release.
  • At least three automations from a customer's roadmap run in production
  • You own the technical side of at least one customer account.
  • The engineers you influence trough your technical leadership build to your standards

Six months in, new automations go from agreed requirement to production in weeks, releases reach every customer from one source, and the team ships faster and with fewer defects because of the standards and tooling you put in place.


About SCC and the CoE

SCC Services Romania is part of SCC, an IT services provider headquartered in the UK. The Agentic AI & Automations CoE works inside SCC's service transformation: customers consume the result as a faster and cheaper IT service, and we agree the automation roadmap with each of them.

Skills

What Lead AI Engineering jobs ask for — and how much of it you have →
Apply

See also

AI Engineering jobs by country — openings, pay and top skills →