Software Engineer

• Build the layer that turns an interpreted instruction into validated platform actions, and the integrations that carry them out. Concord sits above the advertising platforms. A user writes an instruction in natural language, agents interpret it against the account's Advertising Memory, and Concord turns the result into platform actions that are validated, approved where an authorized user is required, and executed through direct API integrations. You will work on that path. It is a distributed systems problem behind a marketing product's surface: third-party APIs with their own rate limits, error semantics and idea of what a campaign is, and no tolerance for a write that should never have left the system. What you will work on
• Direct API integrations with the buying platforms Concord connects to, including DV360, Meta, TikTok, Amazon DSP, Google Ads and The Trade Desk.
• The validation layer between interpretation and execution: permissions, advertiser rules, parameters, formats, quotas and platform limits, all applied before a platform action reaches a platform.
• Advertising Memory: how Skills, which are reusable instructions, and campaign history become account context that later executions can rely on.
• The headless surfaces: the public API, triggers (POST /triggers/...), and askConcord for teams delegating from their own agents.
• Automations that run on a schedule and post their output back to the run.
• The audit history that makes an execution explainable months after it happened. What we look for
• Several years shipping and operating production software, with real ownership of what you built rather than a hand-off at merge.
• Strength in at least one typed language, and a preference for working across the stack instead of behind a single service boundary.
• Fluency with third-party APIs at scale: pagination, rate limits, partial failure, idempotency, and retries that cannot double-spend.
• Judgment about correctness in a system where a bug moves someone else's budget.
• Working English, written and spoken. The product, the code and the documentation are in English.
• Comfort in a small team, where the roadmap changes when a client conversation changes it.

• Experience : in adtech, programmatic buying or ad operations tooling.
• Experience : building on large language models in production, including evaluation and failure handling. - React and TypeScript for product surfaces.
• Experience : with permission models, approval flows or audit requirements in a governed environment.

How we hire

• A first call on the role, the product, and what you are looking for.
• A technical conversation with the engineering team about work you have shipped.
• A practical exercise, reviewed together rather than graded in the dark.
• A final conversation on scope, team and offer.