DysrupIT ir mākoņpakalpojumu un IT risinājumu uzņēmums, kas palīdz organizācijām modernizēt tehnoloģijas un pāriet uz pakalpojuma principā balstītiem darbības modeļiem. Uzņēmuma darbība aptver uzņēmumu mākoņrisinājumus, pārvaldītos pakalpojumus, lietojumprogrammu izstrādi, kiberdrošību, kā arī datu analītiku un informācijas pārvaldību. Tas atbalsta gan mazas un vidējas organizācijas, gan starptautiskus uzņēmumus. Sadarbojoties ar Microsoft un citām mākoņplatformām, DysrupIT apvieno tehnisko izpildi ar talantu attīstīšanu, ieguldījumu kopienā un ilgtermiņa partnerībām, lai veidotu drošas un mērogojamas tehnoloģiju vides, kas atbilst uzņēmējdarbības vajadzībām.
Senior Azure AI Foundry Developer
Build governed, multi-agent AI solutions on Azure AI Foundry, connecting Dataverse and Power Platform for DysrupIT’s Microsoft consultancy client.
Apraksts
- Build multi-agent solutions on Foundry Agent Service with Microsoft Agent Framework, Python, and .NET.
- Develop orchestration, hand-offs, tool calling, and support for long-running tasks.
- Create ingestion, chunking, indexing, and retrieval pipelines with Azure AI Search.
- Restrict permissions and implement grounding and citations so agents use approved sources.
- Package reusable skills and tools, including Foundry toolboxes, MCP servers, and Azure Functions.
- Use Dataverse for system-of-record data and agent state.
- Integrate through the Dataverse Web API, MCP, and custom connectors.
- Trigger Power Automate flows and publish agents in Copilot Studio, Teams, and Microsoft 365 Copilot.
- Set autonomy levels and implement human approvals and audit logging.
- Partner with solution architects and client risk teams on governance.
- Configure tracing, evaluators, regression datasets, and red-team tests.
- Track production quality, latency, and token costs.
- Assign each agent a Microsoft Entra identity and use managed identities and Key Vault.
- Design for private networking, content safety, and data residency.
- Deploy infrastructure as code with Bicep or Terraform.
- Run CI/CD with GitHub Actions or Azure DevOps across development, test, and production.
Prasības
- At least 6 years of software engineering experience, including 2 or more years developing LLM-based applications.
- Practical delivery experience with Azure AI Foundry, including agents, model deployments, Azure AI Search, and evaluation.
- Strong Python and C#/.NET skills.
- Sound engineering practices, including testing, code review, Git, and CI/CD.
- Production experience with agentic approaches such as tool use, planning, memory, multi-agent orchestration, and retrieval-augmented generation (RAG).
- Working knowledge of Azure Functions, Container Apps, API Management, Key Vault, Entra ID, and Application Insights.
- Working knowledge of Dataverse tables, its security model, and Web API.
- Understanding of how Power Platform connects to external services.
- Experience applying guardrails and evaluating AI systems in production.
- Ability to communicate clearly with business and technical audiences.
- Experience developing Copilot Studio agents and publishing them to Teams or Microsoft 365 Copilot.
- Experience building or using MCP servers.
- Experience with Azure AI Document Intelligence, Microsoft Fabric, or Azure AI Content Safety.
- Delivery experience in regulated sectors such as legal, financial services, or government.
- Microsoft certifications such as AI-102, AZ-204, or PL-400 are preferred.
Priekšrocības
- Government-mandated benefits and supplemental HMO coverage.
- A collaborative, professional work environment.
- Career development opportunities in a growing organization.
- Fully remote work arrangement.
- Training and support for career growth.