Senior AI Engineer

➤ Production AI agents using Microsoft Foundry Agent Service: multi-agent workflows, hosted agents, memory, observability, identity integration, secure endpoints, and deployment

➤ Knowledge-grounded experiences using Azure AI Search and retrieval-augmented generation, including vectors, keyword-searchable, hybrid, semantic-ranking and enterprise-permissioned retrieval (When do we need vectorization and when we don’t, when to create handovers between agents, when to rely on the model’s context window, when to finetune and when to use RAG)

➤ Deterministic, source-grounded extraction: pipelines that pull figures and facts straight from documents without hallucinating, attaching citations, provenance and uncertainty flags to every claim so outputs are auditable

➤ Tool-using agents that execute secure actions through Model Context Protocol (MCP), Azure Logic Apps, Azure Functions, and custom APIs

➤ Model-optimized applications using Foundry Models and the model router, balancing quality, latency, and cost

➤ Enterprise-safe systems with prompt shielding, grounded Ness checks, and content-safety enforcement

➤ Reusable reference agents, connectors, and evaluation templates that ship into our UseCaseLibrary and accelerate delivery across the region

➤ Translate business opportunities into robust, production-ready agent architectures

➤ Build pro-code AI agents using Python (FastAPI, Streamlit, Scraping, Serialization, etc..), .NET React.js with the Microsoft Agent Framework (Knowing Langgraph/Langchain/Semantic Kernel is an edge)

➤ Lead the design of multi-agent orchestration and agentic design patterns, including planning, routing, and human-in-the-loop patterns, on the team's most complex builds (understanding distributed systems is a plus)

➤ Review architecture and delivery plans from other AI Engineers before build starts, and act as a technical escalation point during delivery

➤ Design and maintain the delivery lifecycle for agents already in production — evaluation, telemetry, logging, safety, and CI/CD pipelines — using GitHub Actions and pipeline YAML, with infrastructure-as-code (Bicep or Terraform) a plus

➤ Familiarity with networking concepts — VNets, subnets, private endpoints (and private DNS), NSGs, and firewall / allow listing — for deploying AI services securely inside an enterprise tenant.

➤ Familiarity with data engineering and pre-processing — parsing, cleaning, normalization, deduplication and schema creation and engineering/validation (for quantitative and qualitative metrics) — so the data feeding agents and retrieval is clean, typed and reliable.

➤ Understanding model capabilities, benchmarks and limitations — and how to work around them: training-data knowledge cut-off, prompt-injection, context and harness engineering, retrieval grounding, and fine-tuning.

➤ Set and maintain the reference patterns and reusable components other engineers build from

➤ Mentor AI Engineers, including colleagues moving into the role through our internal upskilling path

➤ Support presales and customer proposals with technical input on feasibility and solution design

➤ Significant hands-on experience building production AI agents or AI-infused applications, not just prototypes

➤ Strong software engineering skills in Python or .NET, including modern cloud-integration patterns

➤ Hands-on experience with Microsoft Foundry: projects, SDKs, models, the model router, hosted agents, and Foundry tools

➤ Demonstrable experience building complete agentic workflows in Foundry Agent Service, including tool invocation and deployment

➤ Deep understanding of Azure AI Search and retrieval-augmented generation patterns, including vector, hybrid, and multimodal retrieval

➤ Practical experience implementing AI safety features: prompt shields, grounded Ness checks, and sensitive-content filters

➤ Experience with Azure Logic Apps, Azure Functions, Model Context Protocol (MCP) tools, and secure enterprise integrations

➤ Strong understanding of Microsoft Entra ID and enterprise governance best practice

➤ Experence reviewing other engineers' technical work and giving direct, constructive feedback

➤ Daily proficiency with GitHub Copilot, including effective use of GitHub Copilot Chat in real development work

➤ Comfort with CI/CD pipelines, telemetry, evaluation metrics, and iterative delivery

➤ Backend and streaming services: designing and operating asynchronous APIs (FastAPI/ASGI, or ASP.NET) with streaming responses, middleware, and clean configuration and secrets handling