Senior Founding Engineer – AI Learning Platform

About this role

Location: Hybrid (UK preferred)

Reporting to: Head of Engineering

Key Partners: Kim Faura (Product Lead) & Pravin Paratey (Head of Engineering)

The Split: 80% Deep Building & Coding | 20% Technical Leadership & Team Shielding

ABOUT SALESAPE & ABI

We are building what we believe will become the operating system for millions of small businesses.

Today, we have one main product brand — SalesApe, which helps businesses automate customer conversations, qualify incoming leads, and convert more sales. Alongside this, we are building Self-Serve Abi (Artificial Business Intelligence) — a natural language AI business partner that allows business owners to create, operate, and grow their businesses simply by talking to an AI.

But our long-term vision goes far beyond individual AI agents. We believe the next generation of software will continuously learn from the outcomes it creates. Every customer interaction, recommendation, experiment, and business outcome should make the platform smarter for the next customer. To achieve that, we're looking for a Senior Founding Engineer to build the core intellectual property that ties these products together: our unified, self-improving Learning Platform.

THE MISSION

Your mission is to architect and build the intelligence layer that sits behind both SalesApe and Self-Serve Abi. This platform will capture business events, measure outcomes, identify patterns, and continuously improve the recommendations our AI makes.

Rather than simply orchestrating existing foundational models, you will build a self-improving recommendation and learning engine that compounds over time. Imagine millions of businesses collectively teaching the platform: which sales techniques convert best, which marketing campaigns actually work, and which onboarding journeys reduce churn. Every customer benefits from the learnings generated by every other customer, strictly preserving privacy and security.

This is not a theoretical academic exercise. We will frame your first phase deliberately: prove the learning loop end-to-end on one real, high-value problem—onboarding and retention—using the outcome data we are already generating, from our early trust and engagement signals to the events pipeline and retention dashboard. A concrete early win on a problem we genuinely care about earns us the right, and the real-world data, to tackle the harder architectural questions. From there, you scale and compound the same loop across the rest of the product. This approach drives immediate product value while we build toward the multi-year strategic defensibility moat we need ahead of our Series B.

WHAT THIS ROLE ACTUALLY IS (AND ISN'T)

We are not looking for an "ivory tower" architect or a hands-off engineering manager. We need a highly skilled, pragmatic engineer who is still deeply in love with writing code and shipping systems. We are setting a clear goal—own a learning system that gives our agents opinionated, evidence-backed workflows—and trusting you to own the how. That explicitly includes challenging our assumptions and weighing alternative approaches to get there: we want the strongest architecture, not a predetermined one. The role is split into two primary responsibilities:

- 80% Engineering & Building: You will spend the vast majority of your time architecting, writing, and shipping production-ready code. You will inherit a seeded prototype of our knowledge layer and harden it into a robust, scalable, and resilient production platform.

- 20% Technical Leadership & Shielding: You will partner closely with the Senior Leadership Team to ruthlessly prioritize the technical roadmap. You will guide other engineers on architectural standards and act as a protective buffer—keeping them safe from the daily "noise" of a fast-growing startup so they can focus on deep, uninterrupted builder mode.

WHAT YOU'LL BUILD

You will design, own, and scale the architecture behind a continuously learning platform. Specific areas of focus include:

- Event collection architecture & customer interaction pipelines to capture rich interaction logs cleanly.

- Outcome measurement frameworks to tie AI suggestions to actual business outcomes (sales, retention, clicks).

- Recommendation & feedback loops that let the AI automatically improve its behavioral models based on real evidence.

- Knowledge graphs, vector databases, and memory/retrieval systems that serve as our persistent cross-product intelligence.

- Experimentation infrastructure & feature stores to run secure experiments and manage features efficiently.

- Model independence & deployment strategy to architect the system so we can run our own or open-source models on our own infrastructure where it makes sense—a deliberate lever to stay independent of any single LLM provider on both capability and cost.

- Evaluation frameworks to continuously benchmark and validate prompt and model improvements.

THE HARD PROBLEMS WE'RE BETTING ON

We will be honest with you: this is a high-risk, high-reward bet, and part of what makes it worth doing is that the core problems are genuinely unsolved. The central question you will help us answer is deceptively simple—can we reliably tell whether something our AI did led to a better business result? Getting there means confronting a few hard problems head-on, and we would rather debate them openly with you than pretend they do not exist:

- Causation, not just correlation: knowing what actually worked, and separating the AI's contribution from everything else happening in a business.

- Capturing the outcome: much of the success that matters—a meeting booked, a deal won, a customer retained—happens outside our systems and often isn't tracked today. Instrumenting reliable outcome signals is a first-class part of this role.

- Transfer across businesses: what works for a life-insurance broker probably isn't what works for a roofer. We need to learn which know-how generalises and…