Aumenta le possibilità di ricevere una risposta da questo datore di lavoro
Il tuo curriculum e una lettera di presentazione personalizzati, che corrispondono esattamente a ciò che sta cercando.
Fourthwall is seeking a Senior backend engineer to join an autonomous team working on a Kotlin-based service that handles payments, taxes, and order orchestration across multiple locations. You will own backend features end-to-end, collaborate with product and other engineers, and help scale a system used by hundreds of thousands of creatives and brands.
The role emphasizes reliability, idempotency, and data reconciliation, with a focus on production-readiness and fast iteration using Kotlin,
Creatives, brands and startups run shops, memberships and donations on Fourthwall, on their own domain and with their own data. MKBHD, Y Combinator, ThePrimeagen, Coffeezilla, Acquired, Vox and the Smithsonian are among them. More than 500,000 creatives and companies use it. Investors: Lightspeed, Initialized Capital and Seven Seven Six. Offices in Los Angeles and Poznań.
Our promise is that we will make creating and selling products as easy as having a great idea. To deliver on that promise, our product must be comprehensive, including handling product design, manufacturing, shipping, ecommerce platform, customer support, and taxes.
The only way to do that is with a superstar team who want to work hard, own large areas of responsibility, and who care deeply about making customers happy. And that’s exactly the type of engineers we’re looking for. Smart, motivated, and looking for a chance to make a big impact on a product.
The service you would join has a first commit from 2019. It has run the business since and still goes to production several times a day.
Fourthwall is the merchant of record. We take the payment, handle sales tax and VAT, get the product made in one of 17 locations, ship it and pay the seller. All of that runs through this service: one Kotlin codebase, one deployable. It talks to payment, tax, manufacturing and shipping partners and exposes a public API with webhooks.
A lot of what it does is payments, refunds, payouts and tax. A bug there costs real money, so idempotency and reconciliation matter more here than in a typical CRUD service. The codebase is seven years old and has the usual leftovers. We clean up as we go, but that is not the main job. We are on JDK 26, Kotlin 2.4 and Spring Boot 4.
AI is changing the world unbelievably quickly, and we believe Fourthwall is in an amazing position to be a key part of the new agentic economy. We have an opportunity to become the MCP that agents use to design and order physical products, making it easy for them to connect with world-class manufacturers.
We’ve built a strong foundation to scale from, including having a product our customers love and being profitable, without losing our startup roots and desire to move fast.
Fourthwall is placing bets: features creatives have asked for, agentic interfaces, groundwork for new products, new partner integrations, changes to payments and payouts, security work. Leadership picks them together with the engineers who build them. Teams of one to five form around a project and re-form when it is done. You would be in those discussions and have a say in what gets built.
Before anything gets built we want to know what it would change for users or revenue, and whether there is a cheaper way to find out. Agents with access to all our data surfaces help with the numbers. You make the case to leadership and to the people who talk to customers every day. Sometimes the answer is a smaller version, or no.
Then you build it, schema to rollout flag, and delete the old path once it is live. A week later you check what it changed.
When production breaks, switch the feature off, tell support what to say, size the impact, then fix and reconcile. Support, growth and fraud will ask you questions, and you will have some for them.
The team works through coding agents and builds its own harness: AGENTS.md files, skills, subagents, hooks. You direct the agent, read what it wrote and run the checks. What you merge is yours. When an agent repeats a mistake, fix the skill or the hook rather than the prompt. We pay for the tooling.
We will ask where an agent let you down on production code and what you changed afterwards. If you paste a prompt and merge whatever comes back, or keep agents out of your work, this job will not suit you.
We test all of this in the process: the code in the exercise and the pairing session, the rest in the intro call and the discussion.
Learnable: jOOQ, JobRunr, Pub/Sub, Terraform, MCP tools with Spring AI.
About five hours in total.