AI GTM ENGINEER (H/F)

La mission

About the company : The Revenue Architects is a GTM Engineering Studio that builds modern go-to-marketsystems for the best tech companies in the world. We work with Canva, Descript, andWorkOS -- category-defining B2B companies backed by top-tier investors. We are not a traditional agency. We operate as an embedded engineering team,deploying AI-powered revenue systems that span data, enrichment, outbound, CRM,intent signals, and lifecycle automation. Behind the client work sits our own platform: asuite of internal apps, agents, and data pipelines that let a small team deliver like a largeone. We want to stay small -- a handful of senior people -- and get a 10x out of agentsinstead of headcount. This role is how we get there. Applicants must be eligible for a J-1 ‘Trainee’ visa : o Either be a graduate or awaiting their diploma (diploma must be obtained at least 15 days before departure), AND be able to demonstrate at least 52 weeks of full-time professional experience in a company (including internships) related to the V.I.E. assignment : https://mon-vie-via.businessfrance.fr/destinations/etats-unis The Role : This is a forward-deployed engineer role, pointed at go-to-market. Half of it is client work: you sit with growth and RevOps leaders, understand what theyare trying to move, and ship the system that moves it. The other half is internal: youown the platform and the agent workflows the whole team runs on, and you make themfaster, more reliable, and more reusable every week. The two halves feed each other. What you learn on a client becomes a reusable piece ofthe platform. What you build on the platform makes the next client faster. What Makes a Candidate Stand Out : Software engineering experience followed by a forward-deployed, solutions, or sales engineering role / Has built an agent, eval, or review loop that runs in production -- not just a demo / Has set up CI/CD or automated quality checks on a codebase or data pipeline / Has worked with a warehouse (Snowflake, BigQuery, Postgres) and modeled data for a business use case / Has built Clay workflows, enrichment pipelines, or outbound infrastructure / Has a personal project, side business, or open-source contribution that shows they build on their own / Has worked at a startup or small team where nobody told them what to do next What Gives Us Pause : Pure DevOps or backend background with no customer-facing or business exposure / Pure business or ops background with no exposure to technical problems / Over-scoping and over-engineering: weeks on architecture before anything ships / Cannot point to something they built end-to-end and shipped Must-Have : - AI-assisted building : Claude Code, Cursor, or equivalent; agent SDKs - Engineering : Python or TypeScript, SQL, APIs, webhooks, git, CI/CD - Data : a warehouse (Snowflake, BigQuery, or Postgres) -GTM : Clay or equivalent; Salesforce or HubSpot Nice-to-Have : - Evals and agent quality tooling (Braintrust or similar) - Next.js / Vercel, Supabase - dbt or similar data modeling - Outbound infrastructure (Instantly, Smartlead, HeyReach) What You Will Do : Ship client systems. Take a business objective from a client ("route website visitors to sequences by intent", "fix our account scoring") and turn it into a working system within days, using Clay, the CRM, the warehouse, APIs, and agents. You will be on calls, present your work, and iterate live. Own the internal platform. Maintain and extend TRA's suite of internal apps and the underlying data layer (Snowflake, Supabase) that client work depends on. Decide what the right data model is, what gets automated, and what gets thrown away. Build the quality loop. Every time someone on the team pushes work -- code, a Clay table, a client deliverable -- an agent should review it, check it against how a senior GTM engineer would judge it, and flag when a human needs to look. You build and run that loop: agent reviews, evals, CI/CD for GTM work. Turn one-off delivery into repeatable process. Most clients need the same 80%: get warehouse access, map the CRM, build the data model, wire the signals. You turn what we did ten different ways into one playbook that agents can execute, and you keep it current. Run the research loop. Benchmark new models, tools, and agent frameworks. Test what is coming out, keep what works, write down why. You spend real tokens figuring out how the platform should be built. Raise the team's AI reflexes. Show the rest of the team how to have an agent review their work, how to commit and open PRs, how to get more out of Claude Code. Engineering best practice is part of the job, not a side project. Compensation and Benefits : VIE indemnity per Business France scale (San Francisco rate) Housing supplement Health insurance via Business France Working directly with the founder in San Francisco Access to the entire TRA client portfolio and GTM tool ecosystem Path to full-time employment for outstanding performers