Founding Machine Learning Engineer in London

Hey! We’re Semaloop. Late last year, we stuck a bunch of iPhones in Charlie’s shed, connected them to a computer, and began building. Now, a few months later, we’ve secured over $3m in funding from investors and operators behind companies like Lovable, ElevenLabs, Hugging Face, Linear and Dropbox - and we’re already relied on daily by cutting‑edge companies like Granola and Gopuff.

We saw first‑hand how building apps was being slowed by checks and balances. Whether at early‑stage startups or tech‑giants like Apple (where Rory previously worked), flaky tests blocked releases and QA slowed cycle time. Automating UI testing has been brittle and slow ever since the first smartphone came out in 2007. We’re on a mission to change that by building agents that test apps like humans do: through sight, sound and touch. By solving for the human interfaces, developers don’t have to maintain special APIs, accessibility IDs or integrate new libraries. Instead, teams like Granola just describe how their app should work in plain English. When they ship new builds, Semaloop does the heavy lifting: verifying that the app works, and if not, figuring out what’s broken. That means an end to manual testing, and hours of time back per day for developers.

Crafting a truly magical experience for developers means building some category‑defining technology. That’s why we’re assembling an all‑star team here in London, with the ambition to truly change how software is made. If you’re excited by the idea of teaching computers to understand, reason and interact with digital interfaces, you belong on this mission with us.

The Job

We can’t keep up with the demand for Semaloop and the intelligence layer is at the heart of everything we’re building. As a Founding ML Engineer, you won’t just be training models in isolation. You’ll be building the core systems that allow machines to understand and operate software interfaces like humans do. One day you might be improving real‑time decision‑making in a multimodal agent, the next designing data pipelines from messy real‑world interaction traces. This is not a research‑only role, and it’s not just applied ML either. You’ll own the full loop, from idea, to prototype, to production, to iteration, working directly with real product data and real customer problems. As part of the founding team, you’ll shape how intelligence is built, deployed, and scaled at Semaloop as we grow to our next 100 customers.

Whether you’re training multimodal models, or polishing user experiences, you’ll own how you manage and prioritise your time. In return, we promise to bring together the rest; a world‑class team of incredible people, an endless list of gnarly technical problems and the opportunity to truly achieve your life’s best work.

We’re looking for…

• A Machine Learning Engineer with strong experience in computer vision and an interest in multimodality.
• Hands‑on experience across modern machine learning and classical CV (CNNs, OpenCV).
• Someone who is comfortable taking systems from research to production, showcasing a blend of academic depth (MSc/PhD) and startup execution speed.
• Someone who enjoys prototyping in their spare time, and geeking out over the technical details.
• Hands‑on experience taking ML ideas from "works on my laptop" to real production systems.
• Calmness when life moves quickly and priorities shift.

As a person, you…

• Consider yourself first and foremost, a builder.
• Get excited at the idea of working together in‑person from our office in Old Street (typically at least 3 days a week).
• Love working in a startup environment.
• Obsess over the customer experience, and quality in products you ship.

We know that amazing talent comes from all walks of life. If you find yourself feeling energised reading the above, whether you look like a fit on paper or not, please get in touch!

You can expect to…

• Design and build the core intelligence behind Semaloop.
• Own the end‑to‑end loop from research to prototype to production through to iteration.
• Work across multimodal models powering real‑world computer use and build and integrate multimodal models (vision‑first, expanding to audio).
• Improve performance in the pursuit of real‑time decision‑making.
• Design data pipelines from live product usage (UI interaction traces, failures).
• Train, evaluate, and ship models into production systems.
• Contribute to product architecture, not just models.
• Stay connected with the latest research and advancements in the wider industry, incorporating new technology and ideas into our stack.
• Work on initiatives across the entire business, from your home in engineering, all the way to helping grow our go‑to‑market machine.

What we offer

• Competitive salary + equity
• 28 days holiday + bank holidays
• Private medical & dental (Bupa)
• Generous parental leave
• Pension matched up to 10% (Penfold)
• Life assurance, critical illness & income protection