Senior AI Engineer

As our Senior AI Engineer, you'll own the intelligence layer of the platform — the gateway, tools, retrieval, and evaluation that let an AI assistant read and act inside a regulated ERP as the signed-in user, and never as anything more. In manufacturing that answers to auditors, an assistant that's confidently wrong, or that touches a record it shouldn't, is worse than no assistant at all. Your work is making AI trustworthy enough to put in front of a quality manager.

This isn't a role for working through a queue of well-specified tickets.

 You'll take large features from a loose problem statement to production — agentic workflows, AI on mobile for the shop floor, and the system and infrastructure changes underneath them. You'll decide what to build as much as how, and you'll push for the breakthrough features that make customers switch, not only the ones already on the roadmap.

The ideal candidate thrives in startup ambiguity, moves with real urgency, and has a track record of turning open-ended problems into shipped products without waiting to be told what's next.

What You'll Do

01 — Take large features from idea to production.

Own agentic workflows and AI on mobile end to end. Start from the user's problem, shape the solution, write the plan and the tickets, build the React and mobile interfaces that make AI output trustworthy and actionable, ship it, and measure whether it worked.

02 — Make and own large system changes.

Propose, defend, and carry through the architecture and infrastructure the AI layer depends on — background agents and job queues, observability, cost metering, per-tenant limits, and model version pinning.

Required Qualifications

- 8+ years building and shipping production software, including 2+ years shipping LLM-powered features to real users, and a track record of owning large features end to end.

- AI-native engineering practice — you've rebuilt how you work around agents rather than adding them to an old workflow. Claude Cowork as your primary tool, with strong hands-on experience using Claude Code, Cursor, Codex, or similar; you direct agents through most of the implementation and stay disciplined about the quality, stability, and extensibility of what they produce.

- Product sense and ownership — you work from problems, not specs. You figure out what users and the business actually need, write your own tickets, make the call when the answer isn't obvious, and drive work to done without needing direction repeated.

- Production LLM application engineering — tool calling, structured output, prompt and context design, and streaming, with Claude, OpenAI, or similar, via AWS Bedrock or direct APIs.

- Retrieval-augmented generation in production — chunking, embeddings, and hybrid search on pgvector or equivalent, with access control enforced at retrieval time.

- Evaluation and AI security discipline — you've built eval suites that gate releases, and you design against prompt injection, data leakage through tool output, and cross-tenant access from the start.

- Strong TypeScript, React, and Next.js in production — enough to own the AI experience end to end rather than hand it off.

- Production experience with PostgreSQL — schema design, query optimization, and operating it at scale.

- Hands-on with AWS services — Bedrock, RDS, Fargate, S3, VPCs, IoT Core, DynamoDB, and related infrastructure.

- Disciplined engineering fundamentals — SOLID, DRY, and Domain-Driven Design applied to complex business domains, with unit and integration tests as part of your normal workflow.

- B.S. in Computer Science, Mathematics, Physics, Engineering, or equivalent practical experience.

Preferred Qualifications

- Early-stage startup experience — you've shipped at weekly cadence, owned outcomes rather than tasks, and built something new that changed what the product could do.

- Shipped mobile applications — React Native or similar — for field, operator, or shop-floor use.

- Experience building manufacturing software or Industrial IoT systems — ERP, MES, or SCADA platforms, or machine connectivity over MQTT, OPC-UA, or similar.

- Hands-on with the Model Context Protocol — MCP servers or clients, and OAuth 2.1 authorization for agents.

- C# / .NET experience, as our core services move to a C# and AWS architecture.

- Exposure to compliance frameworks such as SOC 2, GAMP 5, ISO 13485, 21 CFR Part 11, or similar.

Readiness to work from 16:00 till 01:00 in order to have an optimal overlap of work hours with USA based teams.