Head of AI

Overview

Build the AI capability that lets a fast-moving deep-tech company punch above its size — and protect the IP that makes it hard to copy.

Why this role matters

At Swave Photonics we are pioneering holographic chips at the intersection of optics, semiconductors, and systems design. We move quickly and we compete against players many times our size. Doing that well means getting exceptional leverage from every person and every process — and AI is central to how we intend to achieve it.

This role is about the internal side of that ambition: making Swave a company that uses AI as fluently as it engineers hardware, so that our teams build faster and at higher quality. (There is an AI dimension to the product and our research as well, but that sits with other teams — more on scope below.)

That capability is already taking root: our engineers apply machine learning within their own domains — for example, learned surrogates that accelerate electromagnetic simulation by orders of magnitude. Your mandate is to make that way of working the standard across the organisation, not the exception in one corner of it.

We have laid early foundations for an internal AI capability — a structured, self-hosted knowledge base with retrieval and a set of agent workflows on top — and a clear vision for where it should go. Your first mandate is to take it from foundation to a system the whole organisation relies on every day.

This is a builder’s role. You are the person who sees what is possible, earns the trust of engineers and non-technical colleagues alike, and delivers — one concrete, working step at a time.

What you will do

1. 1. Build the substrate (this comes first)

   • Take our internal knowledge-and-agent system from its current foundation to production: a versioned, structured knowledge base — markdown with YAML frontmatter, the pattern Google recently formalised as the Open Knowledge Format (OKF) — with retrieval that agents and people both read and trust.

   • Design and ship agent workflows that do real work: ingestion, synthesis, drafting, plan stress-testing — on infrastructure we control.

   • Architect for data ownership and vendor independence: self-hosted where it protects our IP, with an agent layer decoupled from any single LLM provider — so we can route across multiple models to cross-check and augment one another, swap or upgrade models transparently, and fall back to local models for cost or resilience.

2. 2. Turn it into company velocity

   • Hunt down AI use cases across R&D, operations, and administration — from quick wins to structural transformation — and prioritise ruthlessly by impact and feasibility.

   • Get production workflows live where they are needed, so they compound: faster simulation and design iteration, less manual drudgery, higher-quality output.

   • Build the internal craft — prompting, workflow design, agent orchestration — and the reusable patterns that let others build on what works.

3. 3. Govern it properly (IP first)

   • Own the data-governance posture for AI at Swave: what runs locally versus in the cloud, how partner-confidential and export-sensitive material is handled, and guardrails engineers can actually follow.

   • Keep it lightweight and practical. We care far more about protecting our IP and our partners’ trust than about producing policy documents.

4. 4. Bring the team with you

   • Build a shared baseline of AI fluency and identify ambassadors in each team who carry it forward.

   • You enable others to teach and adopt — you are not here to run a standing circuit of workshops.

A note on scope. There is also an AI dimension to our product and our research, owned by our product and algorithm teams. This role does not own that work — but you will collaborate with the people doing it, so the ability to hold a credible technical conversation with them matters.

Where the role sits

This role reports to general management, a clear signal that this is a strategic priority for the company, not a peripheral function.

You operate across the entire organisation. You are equally at home in a room full of engineers and in a room with people who have never written a line of code.

What you bring

We are looking for someone who thinks strategically and acts pragmatically — and does both with very little need to be the loudest voice in the room.

Must-have

• A real track record building and shipping AI in a live organisation — working systems people actually use, not slideware.

• Hands-on experience architecting AI agents and memory / retrieval systems. You understand RAG architectures in depth