Location: London (hybrid) · Reports to: Founder & CEO
Lua is the Agent OS. We build digital employees — agents that do a real job end to end: they read a company's own systems, make the decision, take the action, and get a little better every time they run. Not chatbots that answer questions — digital workers that do the work, on the company's own data and workflows, and everything they build stays owned by the business. Teams write their logic on Lua (agents, skills, tools, jobs, webhooks, guardrails) and we run the hard parts underneath: infrastructure, multi-provider model orchestration, data, channels (WhatsApp, Slack, Email, Instagram, Messenger, a voice-capable website widget, HTTP API) and monitoring.
Our customers are the companies that run their markets — the operator of one of the world's most-visited heritage sites, a major travel-and-hospitality group, the largest supermarket chain in its country, a listed conglomerate spanning healthcare and consumer goods, large regional banks, social-commerce companies and a microinsurer covering over a million people — across Africa, the Middle East and South Asia. Together they employ 64,000+ people and turn over $5bn+ a year. Backed by Y Combinator, we've raised $5.8M.
Every one of those deals was sold by a founder in a room. It works, and it doesn't compound. Workspaces is how that changes.
We're launching Workspaces: the place where a company's people and its agents do the work together, built around a shared memory of everything the team knows. It is the first Lua product one person can pick up on their own, get value from the same week, and pull three colleagues into the week after. One user becomes a team, a team becomes a company, and the enterprise conversation we run today starts from inside the account instead of outside it.
You own that motion. This is not a brand-and-budget role, and it isn't a marketing function you'll be handed. You own one question: what is the shortest path from a single signup to a team that would notice if Lua disappeared? You'll answer it with experiments, prove it with numbers, and then make it repeat.
Treat the honesty here as part of the offer. We cannot tell you our CAC to LTV. We cannot tell you which channels work for us. There is no instrumented funnel, no agreed activation metric and no attribution in place. You will establish all of it, then spend against it. You are the person running the experiments personally for the first two quarters, and the person hiring the team that runs them after that. When you build that team it will span product, analytics, performance, channel, brand, community and SEO. None of those hires happen before you know which ones the data asks for.
The first user, and the fifth. Own signup → activation → habit → invite → paid, end to end. Define what activation actually means for Workspaces, instrument it, and take responsibility for the number every stage produces.
The loop inside the product. Design the referral, invite and collaboration surfaces that turn one user into a team: shared artefacts, agents worth showing a colleague, network effects that make the next seat obvious. Ship these with product and engineering rather than around them.
Channel prioritisation, from a standing start. Run the portfolio: performance, community, creator and influencer partnerships, content and SEO distribution. Decide what gets funded on CAC, payback and incrementality. Find out where our users already are, get there cheaply, and kill what doesn't pay.
The experimentation engine. Keep a live backlog of a hundred bets and a cadence that clears it: hypothesis, prioritisation, readout, decision. Build a library of learnings so the company gets smarter every month, and make the readouts short enough that people actually read them.
The growth stack, from zero. Stand up event tracking, analytics, attribution, A/B testing, lifecycle and email automation, and experiment tracking. Agree the data model with engineering so every test is measurable on day one. Self-serve dashboards the whole company trusts, owned by you.
The handoff to revenue. Decide when a self-serve account becomes a commercial conversation, and give the founders and the sales team a signal they can act on. Product-led and founder-led are one pipeline, and you own the join.
The growth roadmap. Bring a point of view in week one, sharpen it against the product, and keep it current enough that anyone in the company can read what we're testing this month and why.
Activation rate. Share of signups reaching a first working agent, and how fast that share climbs. This is the one that matters most; the rest are how you get there.
One-to-team expansion. Time from first user to fifth inside an account, and the share of accounts that make the jump at all. This is the whole thesis of Workspaces, expressed as a number.
Loop efficiency. Viral coefficient and cycle time on the referral and invite mechanics you build, measured honestly.
Experiment throughput and hit rate. Experiments live per week, and the share that produce a decision rather than a shrug.
CAC and payback by channel. Baselined in your first quarter, defended every quarter after, with budget moving toward whatever earns it.
We need an operator who is analytical to the bone and happiest with their hands on the funnel. You have built this before, at least once, at a company small enough that you did the work yourself.
ZERO TO ONE, PROVENYou've taken a self-serve product from a few hundred users to a few thousand and built the repeatable motion that got it there. You can walk us through what you tried, what failed, and which two things ended up mattering.
You've worked without a baseline. Landing in a company with no funnel data and no proven channel is the normal state of your career, not a shock.
You'll be more interesting to us if that growth was B2C, prosumer or SMB than if it was mid-market demand generation. Workspaces looks like the former.
You live in the data: SQL-comfortable, fluent in funnel maths, cohort analysis, LTV/CAC and statistical significance. You form opinions from numbers and change them the same way.
You are curious in a way that shows up in your work: you ask the second and third question, and you'd rather find the surprising answer than the flattering one. We index on this heavily.
You've personally run a high-velocity testing programme and can show us the system that sustained it: the backlog, the prioritisation, the readouts.
Practically: we'll ask you early for a growth roadmap and an experiment backlog for Workspaces, built from a product demo and a funnel sketch. The people who do well here enjoy that exercise.
You've shipped growth features with product and engineering, not just campaigns around them. In a product-led motion the product is the channel and you work accordingly.
You've chosen and implemented your own analytics and growth stack. You don't need a data team to get moving.
You set your own targets, hold yourself to them, and report progress with total transparency, including the weeks it went badly.
You're comfortable being the whole function now and a manager of specialists later, and you know which mode you're in.
You've grown an AI, developer-tool or workflow-automation product.
You designed a referral or viral loop that materially moved acquisition, and you can show the maths.
You've grown a product where messaging channels carried real volume: WhatsApp, Instagram DM, Telegram, Slack.
You've hired and managed growth engineers, marketers or analysts.
You can produce or direct creative for paid and social yourself when speed matters.
90 days. Funnel instrumented, activation metric agreed, baseline CAC and conversion established, first wave of experiments shipped and read out to the company.
6 months. A weekly experiment cadence running, one channel and one in-product loop showing compounding returns, and a defensible view of which channels to fund next.
12 months. A growth engine: the loops, the stack, the playbook and the first two hires, converting spend predictably into teams on Workspaces.
You own the entire growth surface of the product that changes how this company sells. No committees, no queue.
You work directly with the founders. Your experiments ship the week you design them.
We are early enough that the playbook is yours to write, and funded well enough that you get to run it. Backed by Y Combinator, $5.8M raised.
Competitive salary and meaningful equity. You're building the engine, so you should own a piece of it.
Send us one thing you've grown: a funnel, a loop, a channel. Tell us the numbers before, the numbers after, and what you'd have done differently. Send your CV alongside it.
See our other open roles at jobs.ashbyhq.com/lua.
