Senior AI Engineer

Job Description

About us

We're a fast-growing digital marketing agency. Just shy of a seven-figure month, sustained for the past four months, with the systems and infrastructure to scale further. ~80-person team across operations, performance, content, and recruitment. The right Senior AI Engineer in this role could realistically 2–3× agency throughput within a six-month window — possibly more — by streamlining operations and shipping the autonomous systems that unblock the next stage of growth.

The mission

Our biggest constraint right now is the gap between what should be automated and what is. We have a pipeline of bottleneck-killing autonomous-system builds queued across every function in the business — recruitment, content production, sales ops, performance reporting, internal QA, customer flows. We need an engineer who can take any of these — voiced as a problem — and come back with a robust, monitored, production AI system that solves it permanently. Not a Zapier zap. Not a POC. A real system.

You'll work directly with our Director on the highest-leverage build projects in the business. The build queue reshuffles week-to-week based on what's bottlenecking us. The role is constant.

The bar is senior. Strong candidates won't just execute the brief — they'll push back on it, tell us what we're missing, and bring tools and techniques we don't currently use. If you want a defined scope and a list of tasks, this isn't the role. If you want ownership over a real engineering footprint with a Director who will back proactive thinking, this is it.

What you'll own

1. The autonomous layer across the business. Anywhere humans are doing repetitive cognitive work that could be carried by a well-built system — recruitment scoring, content production, performance reporting, internal QA, sales operations — is yours to attack. You audit, propose, design, ship. Some of these will be LLM-powered, some classic automation, some hybrid. You pick the right tool.

2. AI integration where it adds genuine leverage. Build LLM-powered tools for screening, scoring, content generation, and operational decision support. Own prompt engineering, evaluation, retrieval, and iteration cycles end-to-end. Integrate AI where it changes outcomes, not for novelty.

3. Architectural voice. Audit our existing systems and identify the highest-leverage upgrades. Document everything cleanly so the team can use, debug, and extend what you ship. Be the technical voice on engineering decisions across the business. Bring perspective from systems you've shipped elsewhere.

4. Real-world example of what's queued right now. To make this concrete: one of the projects in flight is an internal content-production engine that pulls short-form vertical video, regenerates the first frame through a brand-identity-preserving image-generation provider chain, then drives a motion-video provider against the original clip. Stack includes Next.js, Postgres, R2, Apify, WaveSpeed, Nano Banana Pro, Flux Kontext, Kling v3, ffmpeg, worker queue with full audit trail. That's the kind of system we ship — but the role is about being able to ship the next ten of these too, across whatever domain comes up.

To apply

Click Apply. You'll be taken to a short application form. In the form, lead with one production LLM-driven system you have shipped end-to-end — what it does, what corpus or data it learns from or operates against, and what measurable outcome it drove. Generic AI/ML proposals will not be reviewed. We're moving quickly and reviewing applications daily.

Requirements

Hard requirements:

- C2 English — native or near-native fluency, written and spoken

- Proven track record building and shipping LLM-driven systems in production end-to-end (not POCs) — real systems serving real users with measurable outcomes

- Architect-level depth across LLM APIs (Anthropic, OpenAI), retrieval pipelines, evaluation frameworks, structured outputs, cost management

- Confident across the full stack — backend, APIs, databases, frontend integration. You take projects from blank page to deployed product without handoff

- Strong written communicator — clear, structured, no waffle. We hire on written output

Strong signal — stands out:

- You've taken messy, partially-defined business processes and shipped autonomous systems that solved them — and can walk us through the architecture

- You've built systems where "doesn't sound or look AI-generated" was a hard requirement and you actually solved it

- You've architected agentic systems with multi-step decision trees, tool use, autonomous self-correction

- You've solved real failure modes in multi-provider AI pipelines (silent provider degradation, rate-limit cascades, drift detection, intermediate-asset sprawl)

- A portfolio of production work you can walk us through

Mindset:

- Architect, not order-taker. You propose solutions, you don't wait for specs.

- Independent and self-directed. You drive projects from blank page to production.

- High taste. You ship work that's clean, monitored, and built to last — not held together with duct tape.

Benefits

Compensation

- $8,000 – $12,000 USD per month (equivalent to $50–75 USD per hour, ~160 hours/month)

- Top end of range available for architect-tier candidates with proven delivery

- Performance-based bonuses tied to measurable outcomes — revenue lift, hours saved, systems shipped

Working hours / setup

- Remote, full-time, flexible schedule with core overlap with UK working hours (GMT/BST)

- We measure output, not hours