Build intelligence that goes anywhere.

We're a small, focused team working on genuinely hard problems in AI compression and on-device inference. If you want your work to matter — in clinics, in the field, on devices that have never had access to AI — come build with us.

To accelerate the world's transition to sustainable AI.

Sustainable means intelligence that requires dramatically fewer resources — cost-effective, environment-friendly, and able to run in places where the cloud has never reached. The average AI inference wastes 99% of its compute. We're building the architecture that eliminates that waste, one compressed model at a time.

That mission shapes everything about how we work: small team, high ownership, zero theatre. Every line of code, every experiment, every product decision connects back to getting intelligence into the hands of people who need it most.

upto 60x Model size reduction without quality loss
750 MB 7B-quality LLM running fully offline on a smartphone
2 ms Model switching latency for hundred concurrent specialists

A culture built for distributed work — not retrofitted to it.

We didn't go remote because of a pandemic. We went remote because the best people for this work aren't all in one city. Our team spans the UK and India today, with plans to keep growing globally. We communicate asynchronously by default, document decisions in writing, and treat timezone flexibility as a feature — not a workaround.

There are no mandatory stand-ups at inconvenient hours. No "camera on" meetings for the sake of it. You're trusted to work in the way that makes you most effective.

Async by default

Work at your peak — not at someone else's schedule.

No performative presence

Output over optics. We measure what you build, not when you're online.

Written culture

Decisions are documented, context is shared, nobody is left out of the loop.

Annual in-person retreats

We meet in person to build relationships — not to work through a slide deck.

Everything you need. Nothing you don't.

Travel stipend

Annual budget for team meetups and offsites in person.

Wellness allowance

Gym, therapy, meditation — whatever keeps you at your best.

Learning budget

Courses, books, conferences — no approval forms, just learn.

Software & tooling

Full access to the tools you need — we don't nickel-and-dime on software. Plus, a generous amount of Claude credits.

Competitive salary

Benchmarked to top-tier tech salaries in your local market.

Distributed-first culture

No performative presence. We hire globally and work asynchronously.

Three seats. Real impact.

We're a small team hiring carefully. Every person shapes the company.

ML Engineer

Research Full-time Remote

We're compressing intelligence down to what matters. If you find it absurd that a 70B-parameter model needs a server rack to answer a question — and you want to do something about it — this role is for you.

You'll work at the intersection of theoretical insight and production reality: designing compression algorithms, optimising on-device inference kernels, and shipping models that run on the hardware humanity actually has — not the hardware in a cloud data-centre.

Our compression technology achieves upto 60x size reduction while preserving model quality. You'll push that further, contribute to our Fern on-device LLM, and explore world-model architectures that reason causally without requiring internet-scale compute.

What you'll do

  • Research and implement novel neural network compression techniques (beyond quantisation and pruning)
  • Develop and tune inference kernels for CPU, NPU, and mobile GPU targets
  • Benchmark and iterate on Fern — our flagship on-device LLM — across quality and latency metrics
  • Contribute to world-model architecture experiments: causal reasoning, prediction, and planning on-device
  • Collaborate with product engineers to integrate models into the Everyday Series SDK

What we look for

  • Strong fundamentals in ML theory — you think in gradients and loss landscapes
  • Hands-on experience with model optimisation (quantisation, distillation, pruning, or custom methods)
  • Comfort in low-level performance work — profiling, kernel writing, or hardware-aware design
  • Curiosity about causality, world models, or unsupervised representation learning
  • A track record of shipping — papers, open-source models, or production inference pipelines

Stack / tools

PyTorchCUDA / TritonONNX / CoreML / TFLitePythonC++

Product Engineer

Engineering Full-time Remote

Most "AI products" are wrappers around API calls. We're building something different: a runtime for composable AI agents that works completely offline. If the boundary between product, platform, and infrastructure excites you rather than intimidating you, keep reading.

The Everyday Series is our suite of no-code and low-code tools that let individuals and teams build, share, and run AI agents — no cloud account required. You'll shape both the product people see and the infrastructure that makes it possible.

This is a full-stack role, but "full-stack" here means everything from CLI ergonomics to real-time streaming UI to the agent runtime that runs on a device with no internet. You'll have direct influence on what we ship and how users experience AI for the first time without a dependency on the cloud.

What you'll do

  • Build and iterate on the Everyday Series web app, CLI, and agent marketplace
  • Design and implement the agent runtime — sandboxed execution, streaming output, tool calling
  • Create developer-facing APIs and SDKs for building custom agents and integrations
  • Work closely with ML engineers to surface model capabilities through clean product interfaces
  • Own features end-to-end: from scoping and design through to production and user feedback

What we look for

  • Full-stack proficiency — you're as comfortable in a CLI as you are in a design file
  • Strong opinions on developer experience: ergonomic APIs, clear errors, fast feedback loops
  • Experience building real-time or streaming interfaces (WebSockets, SSE, or similar)
  • Ability to navigate ambiguity: this is early-stage product work, not ticket execution
  • Bonus: experience with agent frameworks or local-first architectures

Stack / tools

TypeScriptReactNext.jsNode.jsPython (a plus)WebAssembly (a plus)

Forward Deployed Engineer

Engineering Full-time Remote

Our technology only matters if it runs in the places that need it most — a clinic with no broadband, a submarine, a regulated government network. You're the engineer who gets it there. Part integration specialist, part solution architect, part trusted partner to the customer.

Forward Deployed Engineers at Antelligent sit at the edge between our core platform and the real-world environments where it gets deployed. You'll work directly inside customer environments — from defence labs to hospital networks to industrial facilities — understanding their constraints and doing whatever it takes to make the deployment succeed.

This role is deliberately hard to categorise. On any given week you might be writing a custom integration layer, advising a customer's infrastructure team, debugging a model loading issue on a proprietary OS, or writing the internal runbook that makes the next deployment twice as fast. You'll feed everything you learn back to core engineering so the platform gets better with every deployment.

What you'll do

  • Lead technical onboarding and integration for enterprise and government customers from contract to go-live
  • Embed with customer teams to understand their stack, constraints, and security requirements
  • Build custom integration layers, tooling, and deployment scripts tailored to each environment
  • Diagnose and resolve issues at the intersection of our platform and customer infrastructure
  • Develop reusable deployment playbooks and documentation that scale beyond individual engagements
  • Act as the voice of the customer in engineering discussions — surfacing constraints that shape the roadmap

What we look for

  • Experience deploying software in constrained, air-gapped, or regulated environments (defence, healthcare, finance, or similar)
  • Strong debugging skills across the full stack — you're comfortable without Google and without a stable environment
  • Ability to read a room: you can work with a cautious procurement team one day and a fast-moving ML team the next
  • Systems thinking — you understand how hardware, OS, network, and software constraints interact
  • Bonus: experience with on-device ML runtimes, secure enclaves, or ITAR/classified environments

Stack / tools

PythonDocker / KubernetesLinux / embedded OSREST / gRPCShell scripting

Field Engineer

Sales Full-time Remote

We're pre-commercial on technology that compresses AI models by up to 60× — and we're looking for someone who can walk a CTO through why that matters and close the deal. If you love the moment when a technical customer finally sees the thing you've been building, this is your role.

This is a founding sales hire. You'll own the full commercial pipeline: identifying design partners, running technical evaluations, negotiating pilots, and converting them into long-term contracts. You're selling compression-as-a-service, on-device models, and hardware licensing — to enterprises, research labs, and government bodies who need AI to work where the cloud doesn't reach.

You'll work directly with the founders and engineering team. There's no playbook yet — you'll write it. That means you need to be comfortable in a boardroom and a Jupyter notebook, and know the difference between a quantised model and a distilled one.

What you'll do

  • Identify, qualify, and close deals with enterprise, defence, healthcare, and research customers
  • Run technical discovery calls and POCs — you'll demo the compression service and Fern directly
  • Own the commercial pipeline from outbound prospecting through to signed contract
  • Develop pricing proposals and negotiate licensing agreements with support from founders
  • Feed customer signals back to product and research to sharpen roadmap priorities
  • Build lasting relationships with design partners and become their trusted technical advisor
  • Drive adoption of the Everyday Series among SMB and prosumer customers — from first demo to active deployment

What we look for

  • Proven track record closing B2B deals in deep-tech, infrastructure, or developer tooling
  • Enough technical depth to hold your own in architecture conversations — you don't need to write the models, but you need to understand them
  • Experience running POCs and technical evaluations with engineering buyers
  • Strong commercial instincts: you know when to push, when to listen, and when to walk away
  • Bonus: existing relationships in defence, healthcare, or enterprise AI procurement

Stack / tools

CRM (HubSpot / Salesforce)Jupyter / Python (for demos)Proposal tooling

No forms. No cover letters. Just a conversation.

We think harrowing application forms are a waste of everyone's time. If one of the roles above made you lean forward, we want to hear from you — in your own words.

Drop us a line at [email protected] telling us what excites you about the role and why you'd be a great fit. Attach your CV if you'd like, or share a link to your LinkedIn or GitHub profile — whatever tells your story best.

01

Send us a note

Email [email protected]. Tell us what excites you and why you'd be a great fit.

02

A quick chat

If there's a match, we'll set up a relaxed 30-minute call — no whiteboard, no trick questions.

03

A small task

A short, paid take-home that reflects real work. We respect your time and compensate accordingly.

[email protected]