Coretek is looking for an Agentic AI Developer to build production AI agents for our clients. You'll design and ship systems where an LLM plans, calls tools, and acts against real business systems, then you'll make those systems reliable, evaluated, secure, and affordable enough to run in production. This is a hands-on engineering role in a client-facing consulting environment.

Responsibilities

• Design, build, and ship agentic AI applications on Azure: multi-step reasoning, tool and function calling, retrieval, memory, and human-in-the-loop checkpoints.
• Implement retrieval-augmented generation pipelines, including chunking strategy, embedding and indexing, hybrid and semantic search, reranking, and grounding with citations.
• Integrate agents with client systems through APIs, databases, and Model Context Protocol (MCP) servers, writing the tool definitions and schemas the model depends on.
• Build evaluation harnesses and golden datasets, and treat accuracy, groundedness, and task completion as measured numbers rather than impressions.
• Instrument agents for production: tracing, token and cost telemetry, latency budgets, failure and fallback paths, and alerting on quality regressions.
• Implement guardrails and responsible AI controls, covering prompt injection defense, content filtering, PII handling, output validation, and clear boundaries on what an agent is allowed to do without human approval.
• Tune cost and latency through model selection, prompt caching, context management, batching, and routing simple work to smaller models.
• Apply engineering discipline to AI code: version control, code review, automated testing, CI/CD, and prompt and model versioning.
• Containerize and deploy agent services, owning the build, release, scaling, and runtime configuration of what you ship.
• Set technical direction on client engagements and mentor junior engineers, establishing patterns and standards the wider team can follow.
• Work directly with clients to turn ambiguous business problems into scoped agent use cases, and be honest about what current models can and cannot do reliably.
• Partner with Project Managers, data engineers, and application teams to deliver end to end, and document what you build so others can operate it.

Requirements

• At least 7 years in a software, data, or ML engineering role, with a minimum of 2 years building LLM-based or agentic applications that reached real users.
• Strong hands-on Python development, with real testing, packaging, and code review practice, not scripting alone.
• Practical experience with LLM APIs and agent frameworks, such as Microsoft Agent Framework, Azure AI Foundry, Azure OpenAI, or LangGraph.
• Working knowledge of agent design patterns: tool and function calling, structured output, planning and reflection loops, multi-agent handoffs, and knowing when a deterministic workflow beats an agent.
• Experience building RAG systems with a vector or hybrid search store, such as Azure AI Search, PostgreSQL with pgvector, or Cosmos DB.
• Prompt engineering depth, including system prompt design, few-shot strategy, context window management, and systematic iteration against an eval set.
• Hands-on experience with Docker and Kubernetes, including writing production images, managing configuration and secrets, and deploying and scaling workloads on AKS or equivalent.
• Experience deploying and operating services on Azure, such as AKS, Container Apps, Azure Functions, or App Service.
• Solid API and data fundamentals: REST, async programming, SQL, and JSON schema design.
• Git-based workflow and experience shipping through CI/CD.
• Excellent communication skills, with the ability to articulate complex technical concepts to diverse audiences, including non-technical stakeholders, and to set realistic expectations about AI capability and risk.
• Exceptional analytical and debugging skills, including the ability to diagnose why an agent failed when the failure is non-deterministic and the stack trace is clean.
• Strong knowledge and experience in working with customers in a consultative approach in a technical environment.

Additional Qualifications

• Experience with Model Context Protocol (MCP) server or client development.
• Familiarity with LLM observability and evaluation tooling, such as Azure AI Foundry evaluations, LangSmith, or OpenTelemetry-based tracing.
• Familiarity with Azure networking, identity, and security requirements, including Managed Identity, Key Vault, and Private Endpoints.
• Experience with fine-tuning, distillation, or small language model deployment.
• Experience with Microsoft Fabric, Azure Databricks, or Azure Synapse for the data layer behind AI solutions.
• Experience with document intelligence and multimodal inputs, such as Azure AI Document Intelligence, vision, or speech.
• Experience in a regulated environment (HIPAA, SOC 2, GDPR, DPDP Act) with auditability and data residency requirements.
• Infrastructure as code experience (Terraform, Bicep, or Helm).
• Proven ability to manage multiple client projects and deliver high-quality results on time.
• Experience in Azure DevOps or GitHub for source control and pipelines.