Based in London this company is a high-growth, early-stage startup founded by serial entrepreneurs who have previously built and scaled companies to billion-dollar exits. They develop cutting-edge, cloud-based technology that leverages AI to transform how commercial organisations detect risks and manage real-world events in real-time.
Why them?
High Impact, Early Stage: As an early member of our engineering team, your ideas will directly shape our product roadmap and technical foundations.
Expert Mentorship: Work alongside a small, high-trust, low-ego team of seasoned engineers who have built products at massive scale.
Real-World Engineering: Move beyond toy projects. You will deal with messy, real-world data, building end-to-end features—from prompt engineering and agent design to edge-device optimisation and cloud-scale infrastructure.
Growth Culture: We are obsessed with rapid learning, hands-on development, and fostering an environment where engineers can grow their skills quickly and confidently.
What You’ll Be Doing:
Prompt & Schema Engineering: Building reliable structured outputs from multimodal models.
Agentic Systems: Designing agents that reason, plan, and utilise tools over our data.
Edge & Real-time: Optimising models for performance-constrained edge analytics.
Evaluation: Building robust, automated evaluation pipelines to ensure our AI is actually delivering value, not just "cool demos."
Full-Stack Ownership: Whether it’s backend orchestration, retrieval systems, or occasionally touching firmware/C++, you’ll be a part of the end-to-end journey.
What We’re Looking For:
Commercial experience shipping ML-driven features.
Strong experience with LLM APIs, prompt engineering, and an understanding of how to measure and maintain output quality.
Great instincts: Knowing when to use an LLM, when to use a classical model, and when to use a heuristic.
A "get things done" mindset. We value curiosity and the ability to dig into data when things don't go as planned.
Experience with Go, Python, C++, or frontend frameworks (React/Vue) is a plus, but they prioritise strong fundamentals over specific language expertise.