Founding GTM (London)

Sparkles is building Lovable for large existing codebases.

Everyone is using coding agents now, but many teams are still stuck relying on outsourced development agencies, and it can take weeks to turn an idea into something live in production.

Sparkles is building the AI workflow for teams shipping changes to existing products. We’re making it possible for product, design, marketing, and ops teams to safely make real changes to production codebases through GitHub-native workflows, reviewable pull requests, and the right guardrails.

Sparkles was founded by Dan, who dropped out of University College London and previously worked with the former CTO of Monzo at Iterate.com. We raised $3.5M, backed by Y Combinator, Moonfire Ventures, and angels from OpenAI, Anthropic, a16z, General Catalyst, Cloudflare, and others.

Location: London · In-person · Full-time

With relocation to San Francisco

We're hiring our founding GTM hire — the person who owns how Sparkles gets in front of teams and turns interest into usage and revenue. You'll work side by side with the founders, in person in London, from day one.

This is a builder's growth role, not a manage-an-agency role. You should already use AI tools by default — using Claude, Codex, and agent workflows to run experiments, build landing pages, wire up automations, and move faster than a whole marketing team, without lowering the bar.

• Own the top of the funnel end to end: outbound, content, launches, lifecycle, and paid experiments.
• Talk to users and prospects constantly, find what actually drives adoption, and double down fast.
• Build the growth engine yourself — landing pages, sequences, analytics, and internal tooling — instead of waiting on anyone.
• Run a high volume of experiments, kill what doesn't work, and scale what does.
• Help set the growth bar, the messaging, and the pace of the company.
• This is not a role where you can coast. It requires deep ownership, agency, and responsibility.

• A track record of driving real growth — pipeline, signups, or revenue — ideally at an early-stage startup.
• Roughly 1–5 years is a good baseline, but we'll flex for someone unusually strong.
• Ability to run growth full-stack with high autonomy: strategy, execution, and the tooling to make it happen.
• Strong writing and taste — you can make a page or a cold email that actually converts.
• Comfort with data: you can instrument a funnel, read it, and act on it.
• High agency: you notice problems, propose solutions, and drive things to completion.
• Clear communication and strong judgment.
• Comfort with ambiguity and a founder-speed environment.
• Heavy AI-tool usage as part of your real workflow, not as a novelty.

• Experience marketing to AI Pilled teams and developers, or selling devtools/AI products.
• Experience running founder-led outbound or a PLG motion at a fast startup.
• You can build — comfortable in code, no-code, or with agents to ship your own growth surfaces.
• Experience with GitHub-native, product, or engineering audiences.

Lane 1: Experienced operator — You've already driven growth at a strong startup and want much more ownership, scope, and upside as the person who builds the motion from scratch.

Lane 2: Exceptional emerging builder — Less conventional experience, but you've already run experiments that worked, write unusually well, move fast, and show real evidence of ownership. We care about slope and output over pedigree.

Sparkles is building Lovable for large existing codebases.

Everyone is using coding agents now, but many teams are still stuck relying on outsourced development agencies, and it can take weeks to turn an idea into something live in production.

Sparkles is building the AI workflow for teams shipping changes to existing products. We're making it possible for product, design, marketing, and ops teams to safely make real changes to production codebases through GitHub-native workflows, reviewable pull requests, and the right guardrails.

Sparkles was founded by Dan, who dropped out of University College London and previously worked with the former CTO of Monzo at Iterate.com. We raised $3.5M, backed by Y Combinator, Moonfire Ventures, and angels from OpenAI, Anthropic, a16z, General Catalyst, Cloudflare, and others.

Sparkles is built for real-time AI agent execution at the edge.

Frontend: TanStack Start, React 19, TypeScript, Tailwind CSS 4, Radix UI, TanStack Router

Backend: Cloudflare Workers, Hono, Cloudflare Durable Objects for real-time WebSocket state

Database: Convex for serverless real-time data, plus Durable Object storage for conversation persistence

AI: Opencode running autonomously in isolated sandboxes, routed through a dual-auth relay worker with per-model token and cost tracking

Sandboxing: Daytona cloud sandboxes with snapshot-based fast startup, Docker containers with s6-overlay, and an in-container daemon (Hono + node-pty) that manages Opencode PTY sessions and streams NDJSON events live

Infra: Cloudflare R2, Stripe + Clerk Billing for credit-based usage, and Sentry + PostHog for monitoring and analytics

1. Screening call
2. Work trial with an offer within 24 hours