> Markdown version of [/jobs/ext/3565956-ai-engineer-workbench](https://www.wearedevelopers.com/jobs/ext/3565956-ai-engineer-workbench). Every page supports `.md` or `Accept: text/markdown`. Links point to the HTML versions so they work for humans too. Agent guide: [/agents.md](https://www.wearedevelopers.com/agents.md). --- # AI Engineer, Workbench - **Company:** Brex Inc. - **Location:** San Francisco, CA, United States (Remote available) - **Experience:** Experienced - **Salary:** $216,000.0 - $240,000.0 - **Contract:** Permanent contract - **Skills:** Application Programming Interfaces (APIs), Artificial Intelligence, Applications Architecture, Application Layers, Spreadsheets, Programming Tools, Financial Software, Interaction Design, Software Engineering, Large Language Models, Backend, Build Management, Integration Frameworks, Front End Software Development, Claude, GPT - **Published:** October 3, 2026 - **Apply:** https://www.themuse.com/jobs/brex/ai-engineer-workbench ## About the Role * 5+ years of software development experience delivering customer-facing applications * Experience building or contributing to a production grade agent, agent harness, workbench, or agentic product. * Experience in application architecture and integration design, with the ability to translate interaction needs into extensible interfaces, protocols, and clear boundaries between components. * Deep familiarity with agent clients such as Claude, ChatGPT, and OpenCode, grounded in sustained use. * Strong product judgment and ownership in ambiguous environments. You work effectively across engineering, Product, and Design, challenge decisions constructively, and take responsibility for production quality and customer outcomes., * Full-stack experience spanning rich frontend interactions, backend services, APIs, and data modeling. * Experience building IDEs, developer tools, real-time collaborative applications, or similarly sophisticated web products. * Experience designing extension systems or reusable application platforms adopted by other teams. * Professional experience shipping LLM or agentic products, including human-in-the-loop workflows, tracing, evaluation, or feedback mechanisms. * Experience with financial software or other consequential workflows where transparency, traceability, and user control matter. ## Description As an AI Engineer on our Finance Agents team, you'll build Brex's Agent Workbench: our singular conversational surface where finance professionals work with Brex's agents through documents, charts, spreadsheets, and other interactive surfaces. You'll help customers understand financial activity, investigate findings, revise policies, and steer agents doing consequential work. Your first substantial project will be evolving the existing workbench into an extensible platform that can support multiple agent experiences, including bookkeeping and Brex Assistant. You'll own the extension model, integration protocols, and application architecture that enable other teams to bring their agents and experiences into a coherent workspace. You'll establish the agentic design system with the technical depth to connect interaction design to platform decisions. Working closely with frontend engineers, agent teams, Product, and Design, you'll build across application layers as needed and iterate with customers. The scope includes experiences such as policy editing, reporting, backtesting, and interactive data exploration, with priorities shaped by user needs. Where you'll work This role will be based in our San Francisco office. We are a hybrid environment that combines the energy and connections of being in the office with the benefits and flexibility of working from home. We require three coordinated days in the office per week - Monday, Wednesday, and Thursday. As a perk, we also have up to four weeks per year of fully remote work! Responsibilities * Design and build a reusable agent workbench platform that supports multiple agent personas and enables other teams to integrate their agents and product experiences. * Design and implement extension models, integration protocols, and application architecture that connect agent capabilities with rich, consistent user interactions. * Build and ship customer-facing workflows that combine conversation with documents, charts, spreadsheets, and interactive artifacts, owning UI implementation and the supporting application behavior. * Partner with frontend engineers, agent teams, Product, and Design to turn ambiguous user needs into cohesive experiences, making thoughtful tradeoffs across usability, extensibility, and reliability. * Own features from design through production and iteration: talk directly with customers, investigate failures, and use feedback and observed outcomes to improve how users understand, configure, and steer agents. ## Related Videos - [Livecoding with AI](https://www.wearedevelopers.com/videos/1201-livecoding-with-ai) - [Claude, MCP and n8n, or: How to Automate Your Day with AI-powered Workflows](https://www.wearedevelopers.com/videos/2043-claude-mcp-and-n8n-or-how-to-automate-your-day-with-ai-powered-workflows) - [Developing the Backend with Stefan Lingler, CTO at Shpock](https://www.wearedevelopers.com/videos/100360-developing-the-backend-with-stefan-lingler-cto-at-shpock) - [API, MCP or MCP App? Choosing the right surface for AI agents](https://www.wearedevelopers.com/videos/100305-api-mcp-or-mcp-app-choosing-the-right-surface-for-ai-agents) - [ Evaluating AI models for code comprehension](https://www.wearedevelopers.com/videos/1462-evaluating-ai-models-for-code-comprehension) - [Headless by Design: Building Enterprise Systems That Agents Can Actually Use](https://www.wearedevelopers.com/videos/100092-headless-by-design-building-enterprise-systems-that-agents-can-actually-use) ## Related Articles - [Navigating the AI Shift](https://www.wearedevelopers.com/magazine/629-navigating-the-ai-shift) - [What is Agentic Programming and Why Should Developers Care?](https://www.wearedevelopers.com/magazine/625-what-is-agentic-programming-and-why-should-developers-care) - [ I Gave a Video Editor More Autonomy Than a Trading Bot. On Purpose.](https://www.wearedevelopers.com/magazine/773-i-gave-a-video-editor-more-autonomy-than-a-trading-bot-on-purpose) - [Dev Digest 120 - Apple and peers](https://www.wearedevelopers.com/magazine/455-dev-digest-120-apple-and-peers) - [Dev Digest 132 - Binging WADFlix?](https://www.wearedevelopers.com/magazine/473-dev-digest-132-binging-wadflix) - [How to Become an AI Engineer](https://www.wearedevelopers.com/magazine/331-how-to-become-an-ai-engineer)