Senior Software Engineer, Build and Runtime Reliability

We’re building software and AI agent workflows that span multiple services, repositories, and development environments. We’re looking for an engineer who can take a failing workflow, diagnose it across those boundaries, and deliver a reliable way to run it again. You will own issues from reproduction through repair, verification, and handoff. Success means that another person can install and operate the result, and that the next build or configuration change requires less intervention. This is a hands-on senior individual contributor role combining application engineering, systems debugging, and release engineering. Our current environment includes Python applications, Linux services, Azure infrastructure, databases, external integrations, and AI agent runtimes. **What you’ll do** - Take ownership of failing builds and workflows. Establish the intended outcome, reproduce the failure through the ordinary installed execution path, and identify its cause. - Debug across application code, package installation, service configuration, process identities, permissions, networking, storage, subprocesses, and external dependencies. - Ship maintainable fixes in the component that owns the problem. Coordinate changes and reviews across repositories when a workflow spans several services. - Make setup, installation, launch, and recovery repeatable. Verify that effective runtime configuration matches the intended configuration. - Develop focused regression tests and representative integration checks. Distinguish application failures from faulty tests, incomplete observations, and unavailable external dependencies. - Verify complete workflows, including the actual users or service identities, publication and review handoffs, output readback, and recovery behavior where required. - Preserve useful diagnostics and failure evidence. Make interrupted or failed operations understandable and recoverable. - Deliver concise operating instructions, exact version and configuration references, verification evidence, and clear remaining limitations. - Turn recurring incidents into improvements to existing tools, defaults, checks, and documentation. **What you bring** - Strong software engineering skills, including practical Python experience and the ability to understand and repair unfamiliar application code. - Experience owning software through installation, deployment, operation, and production or customer-facing failures. - A working understanding of Linux processes, environment inheritance, users and permissions, service management, networking, and persistent storage. - Experience with package and dependency management, CI, automated testing, and reproducible builds or deployments. - A systematic debugging approach: form hypotheses, obtain discriminating evidence, revise your explanation, and verify the repair. - Good judgment about scope and testing. You can identify an incorrect checker while preserving the behavior it was intended to protect. - Clear communication about what you observed, what you changed, what actually ran, and what remains unresolved. - Evidence that other people could successfully operate systems you improved. Relevant backgrounds include backend engineering with operational ownership, production engineering, SRE with substantial application development, developer infrastructure, and early-stage engineering roles with broad delivery responsibility. We value demonstrated work over a particular title, certification, or employer. **Useful additional experience** - Cloud resource lifecycle management. - Containers, process isolation, asynchronous workers, queues, and recovery from partial failures. - Database migrations, concurrency issues, authentication, and integrations with third-party services. - AI agent runtimes, model APIs, and workflows whose outputs require both operational verification and substantive review. **How you’ll work** You will work directly with the founder and product engineers. Given an intended workflow, issue history, and constraints, you will assemble the technical context and drive the investigation. You will own closure across component boundaries, with access and review responsibilities agreed at the start. We welcome AI-assisted development. You remain responsible for understanding the changes, verifying their behavior, and producing a result another engineer can maintain. **What success looks like** The initial assignment will focus on one agreed workflow. You will repair it through the supported operating path, reproduce it from a fresh setup, exercise a meaningful configuration variation, and enable another operator to run it from your instructions. Over time, success means fewer recurring failures, faster setup and recovery, fewer avoidable reruns, and less founder or specialist intervention. We will assess completed workflows, reproducibility, and the quality of the handoff alongside the code itself. **What to send** Share your background and one or two examples of systems you made more reliable. Describe the failure, your contribution, how you verified the result, and what another operator could do afterward. Include code or technical writing you are able to share, plus your availability and preferred engagement arrangement.