Member of Technical Staff, Agent Systems

CharacterQuilt is building an AI workforce for enterprise marketing teams.

Humans decide what should exist and approve it. Our agents create the work, operate the customer's real software, and return proof of what happened. We want one person with judgment to be as powerful as an entire agency.

We have raised $13 million and are building a small, deeply technical team in New York City. Everyone works directly with the founder and CEO, close to customers and production. We work together five days a week in Union Square because the product, customer, and engineering loops are still extremely tight.

Enterprise marketing still runs through agencies because launching one campaign means coordinating specialists across Marketo, HubSpot, Meta, LinkedIn, CMSs, and a dozen other systems. It takes weeks.

CharacterQuilt uses computer-use agents to do that work inside the software companies already use. A campaign that took three months can now ship in two days. We raised $13M and were one of the fastest-growing companies in YC's P26 batch.

You'll make the agents themselves much better.

Our agents have to understand a request, use real customer software, create work that is actually good, publish it correctly, and prove what happened. You'll improve that entire loop across browser use, skills, tools, models, and evaluations.

• Build browser agents that can complete long, messy workflows in real enterprise tools.
• Improve skills, prompts, context, tool use, and model routing.
• Turn customer failures into evaluations and durable product changes.
• Build verification so an agent cannot claim success without a real deliverable.
• Work directly with customers, inspect the output, and ship fixes quickly.

• You're a strong software engineer who has built with agents, LLMs, or browser automation.
• You have product taste and care whether the final output is actually useful.
• You like ambiguous problems and fast feedback from real customers.

Five days a week in Union Square. $250K to $350K base plus 0.5% to 3% equity. We do not sponsor US visas.

We build computer-use agents that take enterprise work from intent to production.

The hard part is not teaching a model to click. It is authenticated browser execution, durable state, tenant isolation, skill and context hydration, safe external side effects, exact readback, evaluation, and recovery when models, browsers, or providers fail.

Our stack includes TypeScript and Next.js, Python and FastAPI, Postgres and Supabase, AWS EC2 and S3, browser automation, and frontier models. Engineers work across the product, database, worker, sandbox, browser, integrations, and customer-visible result.

1. A founder conversation about the best agent or product work you have done and what you want next.

2. A focused technical screen using a real agent-quality failure.

3. A deep dive on agent architecture, browser execution, skills, context, evaluations, and product constraints.

4. A paid two-day onsite work trial on a bounded agent and evaluation problem.

5. References and a direct offer conversation.

We will tell you what every stage measures. The work trial is paid and is designed for mutual evaluation. You will see how we build, how quickly customer evidence reaches the product, and what it feels like to work with the founder.