Senior Software Engineer, Agent Platform
Join ShopMy – Powering the Future of Digital Marketing ShopMy is transforming e-commerce by building the infrastructure for human-led curation at scale. We help brands run performance-driven creator programs while enabling top creators to monetize their influence and partner directly with the brands they love—driving discovery through trusted tastemakers and authentic recommendations.ShopMy recently became a unicorn, raising at a $1.5B valuation with backing from Bessemer Venture Partners, Avenir Growth Capital, and Bain Capital Ventures.At ShopMy, you're building the future of human-led commerce for the world's best brands, tastemakers, and shoppers. You move fast and see immediate impact in your work, alongside accomplished teammates who are driven by helping find the world's best things. If you’re excited to help shape the future of curated commerce at the intersection of technology, culture, and taste, we’d love to meet you.About the Role We're looking for a Senior Software Engineer to join our Director of Engineering building ShopMy's agent orchestration platform: the system our AI agents are built, run, managed, and improved on. Your job is the platform itself: the trigger layer that routes every inbound event from Slack, Linear, GitHub, and the ShopMy product; durable multi-step workflows; human-in-the-loop approvals; agent memory; evals; cost attribution; and the console it's all managed through. You'll partner closely with the agent managers who design and run the agents themselves. It's a small team with broad ownership, and the design challenge is keeping it a true platform - model-agnostic, with diverse triggers, composable workflows, and abstractions that make the next agent cheap to launch instead of a one-off build. Your Impact to the Role Design and evolve the orchestration core: a declarative trigger system spanning Slack, Linear, cron, and HTTP; durable workflows; and agent runs that can suspend and resume around human approval Find the abstractions that keep the platform general, so the next agent or workflow composes existing machinery instead of forking a new code path Build the tool layer that gives agents their capabilities: Slack, Linear, GitHub, Notion, MCP, and a dynamic tool registry Build the management console agent managers use to create, evaluate, and improve agents (React) Make agent behavior legible: run tracing, evals and judges, feedback loops, and per-run cost attribution Design and implement APIs in TypeScript and NestJS backed by Postgres Deploy and operate the platform on AWS (ECS, ALB, CloudFormation, CloudWatch) You Are Energized By Platform engineering with your users in the building: agent managers build on what you ship, and feedback arrives in hours, not quarters Working where the patterns aren't established yet: durable execution, human-in-the-loop, memory, and evals for nondeterministic systems are problems you'll help define Finding the abstraction that turns a special case into a capability Solv...