Applied AI Engineer

About SimplyCubed

SimplyCubed is an AI automation business. We deliver custom AI agents trained on a customer's own data and wired into the tools they already run, live in 2 to 5 weeks rather than the six months and $50K+ an enterprise build demands. Our agents take over real, repetitive work across sales, support, and operations: qualifying leads, resolving tickets end to end, and moving data between systems like Slack, HubSpot, Intercom, Zendesk, and Notion. Every agent ships with scoped access, audit logging, and human-in-the-loop controls, a discipline that comes from our founder's background running security for finance and payments companies.

We run a productized model: a repeatable offering delivered the same disciplined way every time, not bespoke project work. We operate as a fully remote, global team of people, specialists, and agents. We value quality over volume, systems that improve themselves over time, and the kind of methodical thinking that catches a problem before it reaches a customer.

The role

We are building the internal agent systems that run our own business and power what we deliver to customers. This role owns the agent workflows themselves, end to end: designing them, building them, measuring whether they truly work, and shipping them into production where real volume and real cost apply.

Where the AI Integration Engineer owns the connective tissue, this role builds the workflow that runs on top of it. Anthropic and many of its peers use the title Applied AI Engineer for exactly this work, so it travels well beyond us.

Work with us at the commitment that fits you, full-time, part-time, or per engagement, always as an independent partner and always fully remote.

What you'll do

• Design, build, evaluate, and ship whole agent workflows in production, not single skills in isolation
• Build the eval frameworks that tell us whether a multi-step agent actually works, judged on the outcome it produces across real cases rather than how confident it sounds
• Own the performance and cost characteristics of those workflows at real volume, and make the tradeoffs that keep them fast and economical
• Work with retrieval over messy, proprietary customer data, getting the right context to an agent without drowning it in noise
• Feed what you learn in the field back into the roadmap, so the product reflects how agents actually behave in production rather than how they behave in a demo
• Partner with the integration and validation roles so a workflow rests on dependable tools and is verified before it ships

Who you are

You are a production engineer who has gone deep on AI systems and brings the judgment that comes with it:

• A strong production engineer with AI-systems depth. You have shipped real systems and you understand how agents behave, fail, and drift once they leave the demo.
• An eval-at-scale mindset. You do not trust a single good run. You trust a measured result across many cases, and you build the harness that produces it.
• Cloud and operations fundamentals. You are comfortable with Docker, CI/CD, an awareness of infrastructure as code, and at least one cloud in real depth.
• Comfortable in the model layer. You can reason about prompts, context, retrieval, and cost as an engineering problem rather than a guessing game.

Nice to have

Hands-on experience with retrieval over messy proprietary data, building eval harnesses for multi-step agents, or running LLM workloads at cost-sensitive volume is a real plus. None of it is a hard requirement if the engineering and measurement instincts are there.

There is a path from this role toward Forward Deployed Engineer for people who want to take that production depth to the front line and own a customer outcome directly.

How we work

• Fully remote, asynchronous-friendly, global
• Direct collaboration with the founder, with real autonomy once direction is set
• Independent partners, not payroll staff, with the same investment in everyone
• Quality over count: a small set of solid, well-measured workflows beats a large pile of fragile ones
• Clear documentation and SOPs as a core deliverable, not an afterthought

Where this role sits

This is one tier of a capability ladder that runs from validating what our agents depend on up to owning a whole customer outcome. We meet you at the tier that fits you now and grow with you as the work and your capability line up.

1. 1 Agent Skills Engineer Validate and harden the skills, tools, and integrations our agents rely on.
2. 2 AI Integration Engineer Build and own the integration layer those agents run on.
3. 3 Applied AI Engineer You are here Design, evaluate, and ship whole agent workflows in production.
4. 4 Forward Deployed Engineer Own a customer's outcome end to end, embedded in their context.

To apply

Tell us about an agent or AI workflow you shipped to production: how you knew it actually worked, and what it took to keep it fast and reliable. We care more about how you think than about a polished résumé.