AI Engineer (Python and/or TypeScript)

About the client
Our client is a strategy-led digital transformation firm focused on composable commerce and intelligent automation. We work with mid-market and enterprise organizations on high-stakes technology decisions: platform evaluation, roadmap development, implementation, and ongoing optimization.
We’re not chasing trends. We build agentic systems that hold up in production, and we stay accountable well past launch.

The role
We’re hiring an AI Engineer to design, build, and harden agentic AI systems for our clients’ commerce and order operations workflows. You’ll work across our Agentic AI practice, translating client business problems into production-grade agents, RAG pipelines, evaluation harnesses, and integrations with the broader composable commerce stack (commercetools, Algolia, Fluent Commerce, Contentstack, and others).
This is a hands-on engineering role with direct client exposure. You’ll own technical execution from assessment deliverables through implementation and post-launch optimization. You’ll write production code, ship agents to client platforms, and stand behind what you build when it goes live.
This role requires someone who can operate independently in client-facing situations and contribute to how we build this practice, not just execute against a defined playbook. If you need close direction to do your best work, this isn’t the right fit.
The immediate engagement this role is hired into involves building an agentic system on Microsoft Foundry, so familiarity with that platform is a real advantage coming in.

What you’ll do

• Build production-grade agents using the tools and platforms relevant to each engagement, including MCP server integrations, A2A endpoints, and connections to client systems (commercetools, OMS, search, CMS).
• Design RAG and grounding pipelines with attention to retrieval quality, citation integrity, and cost.
• Implement evaluation and observability: build eval sets, establish tracing, and create the feedback loops that let us prove and improve agent quality over time.
• Write production Python and TypeScript for agent orchestration, tool implementations, data transformation, and integration with client commerce platforms and APIs.
• Support strategy and assessment engagements led by senior Aries practitioners, contributing technical depth to client-facing deliverables.
• Collaborate with Architecture, Commerce, and Order Operations practices to design agentic solutions that fit cleanly into MACH-aligned, API-first systems.
• Stay involved post-launch. Tune prompts, expand tool coverage, manage memory and context, and govern agent behavior as client volume scales.

What you bring
Required

• 4+ years of professional software engineering, with at least 2 years building production AI/LLM applications, not just experiments.
• Strong primary fluency in Python or TypeScript, with working proficiency in the other.
• Production experience building agents on at least one major platform: Microsoft Foundry, Azure AI Studio, AWS Bedrock, Google Vertex, or OpenAI direct. You’ve built agents, deployed models from a catalog, and integrated tools.
• Practical knowledge of agentic patterns: tool/function calling, multi-agent orchestration, memory, RAG, grounding, and the failure modes each introduces.
• MCP fluency. You’ve built or consumed MCP servers in a production context, not just followed a tutorial.
• Experience building eval sets, running them against agent outputs, and using results to drive iteration.
• Comfort working directly with client engineering and business stakeholders. You can defend a technical decision on a call without hiding behind jargon.

Strongly preferred
Prior consulting or client-services experience. This role has a real client-facing dimension; people who’ve worked in that environment tend to ramp faster on the parts that aren’t purely technical.

• Microsoft Foundry experience specifically. Our current client engagements run on Foundry, and hands-on familiarity with the Foundry Agent Service, Foundry IQ, and the azure-ai-projects SDK is a meaningful advantage.
• Azure fluency: Entra ID auth, RBAC, networking basics, App Service or Container Apps, Key Vault, and the cost mechanics of running models at scale.
• Understanding of MACH and composable architecture. You know why API-first matters and can design agents that fit cleanly with the rest of a stack.
• Production experience on more than one major model provider, so you can speak to trade-offs from experience rather than speculation.

Nice to have

• Experience with commerce or order operations systems: commercetools, BigCommerce, Shopify, an OMS (Fluent, Manhattan, IBM Sterling), or a major search platform (Algolia, Constructor, Bloomreach).
• Experience with Voice Live, real-time agent runtimes, or multimodal agents.
• Contributions to open-source AI tooling or MCP servers.
• Experience with fine-tuning, distillation, or bring-your-own-weights deployments.

How we work
Remote-first, with the potential for occasional client travel in North America. This is a strategy-led role, not staff augmentation. You’ll be the technical lead on engagements, which means real judgment, real accountability, and real scope ownership. Fixed-scope assessments are part of our model: when we commit to a client, we mean it.
Our AI and Agentic practice is actively being built. The person we hire will have a hand in shaping it.