Senior / Staff Software Engineer, AI Systems

Recruiting From Scratch•San Francisco, CA

5d•Hybrid

About The Position

Our client is an early-stage, VC-backed technology company building AI-powered systems that automate complex workflows in a highly regulated domain. The engineering team is small and closely connected to product, operations, and customers. Engineers work directly with domain experts to understand nuanced workflows, turn them into reliable software systems, and iterate quickly based on real-world usage. The company is building sophisticated AI agents and infrastructure rather than simple chatbot experiences. The engineering challenges span agentic workflows, backend services, workflow orchestration, evaluation infrastructure, event-driven systems, observability, and applicant-facing interfaces. This is an opportunity for a senior or staff-level engineer to have substantial ownership over foundational AI systems and backend architecture in a high-agency startup environment.

Requirements

4+ years of experience shipping production software. 5+ years preferred; candidates with 7–9 years can be particularly strong fits.
Deep experience building backend systems and production infrastructure.
Experience building AI/ML-powered applications.
Strong experience owning projects end-to-end from ambiguous requirements through deployment and ongoing operation.
Experience working at a VC-backed startup or as an early employee at a growth-stage company.
Demonstrated ability to operate with high autonomy in small engineering teams.
Experience shipping durable, production-grade systems rather than primarily conducting experimentation or research.
Deep expertise in TypeScript and Python.
Strong backend engineering fundamentals.
Experience building production APIs.
Experience with complex systems and backend architecture.
Strong understanding of LLM-powered applications.
Experience building with LLMs beyond basic chat interfaces.
Understanding of tool use, context management, evaluation, and unreliable model outputs.
Strong understanding of workflow orchestration and asynchronous systems.
Experience with queues, events, retries, idempotency, and failure recovery.
Experience with relational databases, particularly PostgreSQL.
Experience building production AI/ML-powered applications.
Experience designing or operating agentic systems.
Understanding of AI evaluation and quality measurement.
Ability to build systems that monitor and evaluate AI outputs.
Experience handling unreliable or nondeterministic AI behavior.
Understanding of agent observability and debugging.
High agency and strong independent judgment.
Comfortable working through ambiguous problems.
Strong ability to research unfamiliar technologies and domains.
Comfortable prototyping and validating solutions quickly.
Strong product intuition and understanding of how systems are used by customers.
Excellent communication and collaboration skills.
Comfortable working directly with domain experts and non-engineering teams.
Strong engineering judgment and attention to system quality.
Comfortable working in a small, fast-moving startup environment.

Nice To Haves

Familiarity with FastAPI or Django is a plus.
Experience with React and modern frontend systems is a plus.
Familiarity with GraphQL is a plus.
Experience with complex document workflows or domain-specific automation is a plus.
Experience operating agents in production is particularly relevant.

Responsibilities

Build Production Agentic Systems: Build AI agents that automate complex, multi-step workflows with appropriate human review and intervention. Design systems that use LLMs beyond basic chat interfaces, including tool use, context management, evaluation, and handling unreliable outputs. Turn early AI experiments into reliable, production-grade systems. Work closely with domain experts to understand nuanced processes and translate them into software behavior. Build systems that balance quality, reliability, latency, cost, and user experience.
Design Workflow & Backend Infrastructure: Redesign and improve workflow orchestration for long-running processes, dependencies, retries, and failure recovery. Build and maintain backend services, event pipelines, and scheduled job infrastructure. Design asynchronous systems using queues, events, retries, idempotency, and recovery mechanisms. Improve system architecture and data models as the product scales. Build APIs and backend infrastructure that support complex end-to-end workflows.
Build Evaluation & Observability Systems: Build evaluation infrastructure to measure AI output quality and identify regressions. Develop systems that help improve agents, prompts, and models through structured evaluation. Make agent execution observable so teams can understand system behavior and debug failures. Track and improve quality, latency, reliability, and cost. Build tooling that allows internal teams to monitor and supervise AI workflows.
Own Products End-to-End: Lead projects from ambiguous requirements through design, implementation, deployment, and ongoing operation. Work directly with product, legal/operations, and customer-facing teams to understand real-world problems. Connect AI and backend systems to internal tools and user-facing interfaces. Independently research unfamiliar technical problems, prototype solutions, and validate approaches with users. Make strong engineering decisions without requiring extensive direction or predefined processes.