Principal Agentic Commerce Engineer

AI agents are starting to shop for people. They search catalogs, compare products and check out through new protocols: UCP for commerce, A2A for agent-to-agent, AP2 for payments. Almost every store can now be found by an agent. Very few get chosen, and only a small share of agent shopping sessions end in a purchase.

Speedway is ACOMM’s answer: a platform built around UCP that gets retailers picked by buying agents and wins the sale. We’re looking for the engineer who builds it.

You’ll own Speedway’s engineering end to end, from the first commit and architecture to production.

In your first 90 days: you’ll be onboarded to the Speedway product vision, and you’ll work across all of our core product. We’re moving quickly, and you’ll ship code from the start.

In your first year:

• Connect Speedway to retailers’ commerce platforms, inventory, payment providers and fulfillment, with reusable connectors rather than one-off builds.
• Keep it fast. Agents won’t wait, so we design for agent-ready response times.
• Build security in from the start: verified buying agents, defenses against prompt injection through product content, and protection against refund abuse.
• Instrument everything. Every query, answer and order feeds the learning loop that makes Speedway better, and the attribution our customers rely on.
• Track the standards as they move (UCP releases, A2A, AP2, card-network agent protocols) and help shape where they go.

Google Cloud. Gemini Enterprise Agent Platform: Agent Development Kit (ADK), Agent Engine, Agent Studio. Gemini models. UCP, A2A, AP2 and MCP. We’ll choose the rest of the stack with you.

• Google Cloud in production. You’ve designed, shipped and run real systems on it.
• Gemini Enterprise Agent Platform. Hands-on agent work with it or its Vertex AI predecessors (Agent Builder, ADK, Agent Engine).
• Agent protocols and payments. Experience with A2A and AP2, and with payments: payment providers, tokenization, PCI-DSS scope, disputes.
• AI engineering. Production LLM and agent systems: tool use, evaluations, guardrails, latency and cost.
• A degree in computer science or a related field, or equivalent experience building systems at this level.

• E-commerce platforms such as Salesforce Commerce Cloud, commercetools, Adobe Commerce or Shopify.
• Identity standards: OAuth 2.0, OpenID Connect, HTTP message signatures, agent verification.
• Contributions to open protocols or open-source agent tooling.
• Early-stage or founding-engineer experience.

• Industry-leading work in a field that changes every month. Agentic commerce is being defined right now, and you’ll be building at the front of it, not maintaining yesterday’s stack.
• Real ownership. At a big company you’d own one slice of this. Here you own the whole system.
• Remote-first. Build with agents, work from anywhere.
• Big ideas come from anywhere. We’re a small, human-centered team of designers, engineers and data scientists, and the best idea wins.

Two interviews, three at most.

1. Technical conversation with our CTO (60 minutes): a system you’ve built, and how you’d approach a UCP storefront and selling agent.
2. Design session with the team (90 minutes): we sketch part of Speedway together. No trick questions, and no unpaid take-home project.
3. If needed, a final conversation with the founders. Then references.

Applications close 31 January 2027. Email careers [at] acomm [dot] ai with your résumé or LinkedIn, and a few lines about something you’ve built that you’re proud of.

ACOMM is an equal opportunity employer. We welcome applicants of every background, and we’ll make reasonable adjustments to our process if you need them.