Senior Software Engineer (Agents)

Recruiting From Scratch • San Francisco, CA

2d • Onsite

About The Position

Recruiting from Scratch is hiring a Senior Software Engineer for an early-stage AI company in San Francisco building intelligent software to automate complex engineering workflows. The platform uses AI agents to operate professional engineering applications, helping teams design and develop complex physical products. This is an on-site role, five days per week, on a small, highly technical team where engineers have significant ownership over core product development.

The engineer will build and own core agent-system components such as tool calling, state and context management, error recovery, and model routing. They will design reliable systems for executing complex multi-step workflows across desktop engineering applications, optimize latency, reliability, token efficiency, and cost, and develop production systems that work with real desktop applications in secure or restricted enterprise environments. The role includes ownership of agent infrastructure from design through implementation and production deployment.

The position also contributes to evaluation infrastructure and benchmarks grounded in real engineering workflows. The engineer will work with domain experts and user researchers to turn user requirements into measurable evaluation cases, identify capability gaps, and improve end-to-end workflow coverage. Success is measured through task completion, workflow coverage, reliability, and token efficiency.

The engineer will write maintainable production code, contribute to architectural decisions and code reviews, and collaborate with AI engineers, product, integration teams, researchers, domain experts, and customers. Prior experience building LLM agents is not required; strong software engineering fundamentals and an interest in agentic systems are valued. Compensation is a base salary of $180,000–$330,000 plus competitive equity with a four-year vesting schedule. Visa transfers and new sponsorships, including OPT, H-1B, and TN, are available.

Requirements

• 3–10 years of professional software engineering experience building, shipping, and maintaining production systems.
• Strong computer science fundamentals and sound software engineering judgment.
• Production experience with Python or another modern programming language.
• Experience designing and maintaining production systems.
• Experience with backend systems, APIs, state management, and system reliability.
• Ability to debug complex issues and reason through failure scenarios.
• Ability to own technical problems end to end, from requirements through implementation, testing, and production reliability.
• Ability to reason about multi-step workflows, state management, failure handling, concurrency, and system performance.
• Good engineering judgment and code quality.
• Comfort working in a fast-paced, early-stage startup with significant individual ownership.
• Strong ownership mentality and ability to work independently.
• Comfort working across technical and product boundaries and in an in-person team environment.

Nice To Haves

• Experience with LLMs, AI agents, or model-powered applications.
• Familiarity with agent observability and evaluation tools such as LangSmith or Logfire.
• Experience with CAD, CAE, PLM, or other complex desktop engineering applications.
• Experience with COM or desktop application integrations.
• Exposure to enterprise environments, locked-down workstations, or on-premise deployments.
• Experience building reliable systems across multiple tools and long-running workflows.
• Experience in an early-stage startup or small, high-ownership engineering team.
• Experience in agentic systems, tool calling, orchestration, distributed systems, workflow execution, context management, model routing, evaluation frameworks, testing infrastructure, benchmarking, performance optimization, latency reduction, cost efficiency, desktop application automation, or integrations with complex software.

Responsibilities

• Build and own agent-system components, including tool calling, state management, context management, error recovery, and model routing.
• Design reliable systems to execute complex, multi-step workflows across desktop engineering applications.
• Optimize agent performance for latency, reliability, token efficiency, and cost, including per-workflow budgets.
• Build production systems that interact with desktop applications in secure and restricted enterprise environments.
• Own agent infrastructure from design through implementation and production deployment.
• Help design and develop evaluation infrastructure for measuring agent performance on real-world engineering workflows.
• Build reproducible, challenging benchmarks based on actual user requirements.
• Work with domain experts and user researchers to understand engineering workflows and translate them into measurable evaluation cases.
• Identify agent capability gaps and improve the proportion of workflows the system completes end to end.
• Measure outcomes through task completion rates, workflow coverage, reliability, and token efficiency.
• Write high-quality, maintainable production code and contribute to architectural decisions.
• Participate in code reviews and help establish engineering standards.
• Collaborate with AI engineers, product, integration teams, researchers, and domain experts.
• Work with customers during proof-of-concept engagements to understand real-world usage and improve agent behavior.
• Help shape the technical direction of the core agent platform.

Benefits

• Competitive equity with a four-year vesting schedule.
• Opportunity to work on a technically challenging product at the intersection of AI and engineering software.
• Significant ownership over core systems and product direction.
• Work on core product systems that determine product capabilities and value.
• Build agents for real engineering software and enterprise environments.
• Work directly with senior technical leadership.
• Contribute to a new category of reliable AI systems for complex engineering workflows.
• Join a small, ambitious, in-person team where engineering decisions directly affect the product.

What This Job Offers

• Job Type: Full-time
• Career Level: Senior
• Education Level: No Education Listed