Job Description: AI Product Engineer
Location: San Francisco, CA (Hybrid — In-office Tuesday, Wednesday, Thursday) Compensation: $180,000 – $220,000 base + equity + benefits
Reports to: VP of Product/Engineering
About Smart Access
Smart Access is the AI execution layer for supply chains. We're a Frontline Execution Platform that helps the world's largest warehouses, manufacturers, and logistics operations close the gap between defined standards and actual frontline behavior. Our system connects standards, observations, coaching, and intelligence into a single execution loop that drives consistent, measurable performance on the floor.
We work with operational leaders, frontline supervisors, safety teams, and continuous improvement leaders to systematically close the execution gap — the distance between what standards say should happen and what actually happens. As supply chains adopt AI, Smart Access is positioned to become the operational intelligence layer that AI agents rely on to drive frontline action.
The Company just closed a Series A funding round, marking a significant milestone in its growth journey. This is a pivotal moment to join; the team is lean, the trajectory is steep, and the roles being hired now will be foundational to how Smart Access scales. If you thrive in a high-ownership, high-impact environment and want to help build something from the ground up, this is the opportunity
The Role
We're hiring an AI Product Engineer to build Smart Access's MCP server and the agentic workflows that sit on top of it. This is a defining technical role at the company — you'll architect how AI agents securely access, reason over, and take action within the Smart Access platform.
This role reports jointly to the VP of Product and the Head of Engineering. You'll partner closely with both — shaping product direction with the VP of Product and owning technical architecture and delivery with the Head of Engineering. The dual reporting reflects how foundational this work is: it sits at the intersection of product and engineering execution.
You'll start by building our MCP server: the layer that exposes Smart Access capabilities (SOP retrieval, observations, action plans, skill building, certifications, intelligence) to AI assistants like Claude, ChatGPT, and Copilot — and to tailored agents we'll build for our customers. From there, you'll build the agentic workflows that let frontline leaders, operational leaders, and functional stakeholders direct AI to perform real work in our platform.
This is a build-from-scratch role with significant ownership. You'll set the technical direction, make foundational architecture decisions, and ship capabilities that go directly into the hands of operational leaders running real supply chains.
What You'll Do
- Design and build the Smart Access MCP server — including authentication, OAuth scopes, tool definitions, resource exposure, and audit logging
- Build agentic workflows that allow customers to direct AI to take action in the platform (observations, action plans, skill assignments, certifications) with appropriate human-in-the-loop guardrails
- Architect the safety and approval layer — define how sensitive actions are scoped, approved, and audited so that customers trust AI to operate inside their workflows
- Develop tool schemas and prompts that LLMs can reliably reason over, balancing flexibility with deterministic execution
- Partner directly with the VP of Product on roadmap, prioritization, customer use cases, and which capabilities to expose
- Partner directly with the Head of Engineering on architecture, technical standards, hiring, and engineering execution
- Build evaluation infrastructure to measure agent quality, tool-call accuracy, and task success rates
- Ship customer-facing AI experiences that bring agentic capabilities to frontline associates and operational leaders
- Set engineering standards for how Smart Access builds AI-native features going forward
What We're Looking ForRequired
- Minimum 6 years of software engineering experience, with strong backend or full-stack capabilities
- At least 1 year of hands-on experience building MCP servers — you've designed tool schemas, implemented OAuth flows, and shipped MCP capabilities to production or production-adjacent environments
- Experience building AI agents or agentic workflows — you understand the difference between LLM features and agent systems, and you've worked on the latter
- Proficiency with modern AI APIs and frameworks (Anthropic, OpenAI, LangGraph, Vercel AI SDK, or equivalents)
- Strong systems thinking — you can design APIs, data models, and permission systems that scale across enterprise customers
- Production-grade engineering rigor — testing, observability, security, and incident response are habits, not afterthoughts
- Excellent written communication — you can explain technical decisions clearly to product, design, and customers
- Comfort working with multiple senior stakeholders — you can manage priorities across both product and engineering leadership without needing constant alignment
Nice to Have
- Experience in supply chain, logistics, manufacturing, or frontline operations software
- Background in enterprise B2B SaaS, especially with multi-tenant architectures
- Experience with OAuth 2.1, SSO/SAML, and enterprise authentication patterns
- Prior work on RAG systems, semantic search, or document retrieval at scale
- Experience evaluating LLM systems (eval pipelines, benchmark design, hallucination measurement)
- Familiarity with WMS, HRIS, or operational system integrations
How You Work
- You ship. You're biased toward shipping working software over perfecting designs.
- You're comfortable with ambiguity and excited by greenfield problems.
- You think about the customer's customer — the frontline worker who'll actually use what you build.
- You take security and safety seriously, especially when AI is taking action in real systems.
- You're a strong collaborator who can navigate product and engineering priorities simultaneously.
- You’re scrappy. You find ways to make progress with limited resources, prototype before you have permission, and turn constraints into creative solutions. At an early-stage startup, ambiguity and resource limits aren't obstacles.
Logistics
- Location: San Francisco office, hybrid in-office Tuesday, Wednesday, and Thursday
- Compensation: $180,000 – $220,000 base salary, depending on experience, plus equity and benefits
- Start date: As soon as possible
