Browser Agent Engineer

Browser Agent Engineer About the Role Today, 99% of the world's information is only accessible through a browser. Yet browsers were never designed for AI Agents-they were built for humans. We're changing that. Our vision is to make AI Agents first-class citizens of the web , with their own identity, inspectable memory, delegated permissions, and replayable behaviors. The browser should no longer be merely an automation tool-it should become the place where Agents truly live and work. The mission of this role is to build the runtime foundation for that world, enabling Agents to reliably complete long-running tasks in the unpredictable, messy reality of the web-from launch to archived results-with minimal latency and no human supervision. Core Responsibilities Browser Runtime Build the core browser runtime, including: Session isolation Browser profile management Screenshot and screen recording Live browser view Persistent authentication sessions Crash recovery Long-Running Task Orchestration Design reliable execution systems for long-running Agent tasks, including: Task and session state machines Idempotent retries Cancellation handling Resume-from-checkpoint execution Automatic resource cleanup Agent-Driven Web Navigation & Interaction This is not traditional web scraping. Your mission is to enable Agents to understand and interact with websites the way humans do.

Responsibilities include:

Extracting and structuring information from dynamic web pages without relying on fixed CSS selectors, using semantic understanding instead. Automating multi-step forms and authentication workflows.

Handling complex interactions such as:

Pagination Infinite scrolling Pop-up dialogs Dynamic UI components Adapting strategies in anti-bot environments through: User-Agent rotation Human-like interaction patterns Intelligent waiting strategies Detecting and reporting CAPTCHA challenges and mandatory verification flows. Producing structured outputs that downstream systems can consume directly. Agent Loop & Tool System Design and improve: Task planning Tool invocation Structured outputs Failure attribution Multi-Agent orchestration Developer APIs Build production-grade developer interfaces, including: REST APIs Server-Sent Events (SSE)

SDKs Ensure:

Clear and actionable error codes Reliable artifact downloads Stable developer experience LLM Infrastructure Own the AI invocation pipeline, including: Prompt version management Model routing Cost tracking End-to-end request tracing Observability Build comprehensive observability systems that enable engineers to reconstruct production failures within minutes using: Logs Distributed traces Event streams Browser screenshots Execution recordings What We're Looking For Required Qualifications Browser Automation Expertise Deep understanding of browser automation technologies, including: Chrome DevTools Protocol (CDP) Playwright Puppeteer Remote browsers Dynamic rendering Anti-bot detection and failure patterns LLM Agent Engineering Strong understanding of modern AI Agent architectures, including: Agent loops Tool calling Context window optimization Prompt version management Long-Running Workflow Systems Experience designing reliable long-running task infrastructure, including: State machines Worker systems Idempotency Timeout management Cancellation Failure recovery Backend Engineering Strong Python expertise, including: Asynchronous programming FastAPI Pydantic SQLAlchemy pytest Event-Driven API Design Experience designing production APIs using: Server-Sent Events (SSE) OpenAPI Stable SDK contracts Actionable error handling Observability Hands-on experience with production observability platforms such as: OpenTelemetry Prometheus Structured logging Experience 5+ years of backend engineering, browser automation platforms, or AI Agent development. Proven experience delivering production-grade browser automation systems or long-running workflow platforms. Preferred Qualifications Candidates with one or more of the following experiences will stand out: Engineering experience with browser-native Agent platforms such as: browser-use Browserbase Stagehand Skyvern or similar products Hands-on experience with: Human-in-the-loop workflows Authentication/session management Anti-bot mitigation Failure classification for dynamic web interactions Experience with distributed workflow platforms such as: Temporal Celery Kafka Redis Queue Maintains production-grade open-source automation tools, web crawling frameworks, or AI Agent projects with active real-world users.