AI Lead - Innovation

About RUSH

RUSH is an integrated design and technology studio. We design and build software with enterprise clients. Delivery sits with mixed teams of engineers, designers, and strategists.

Values: Put people first. Push the boundaries. Play the long game. Say it like it is. Live for the RUSH.

The AI Chapter runs two key service lines. This role leads Innovation: a small team that sits with a client, keeps finding problems worth solving, and proves them with working software.

About the role

You support and deliver the AI Innovation service line as a Tech Lead who also does architecture and pre-sales.

Typical work is a problem the client cannot quite solve with a chatbot and a spreadsheet. Sometimes that is inside the organisation: reviewing documents, checking photographs, triaging cases. Sometimes it is customer-facing: helping someone find the right product, answer, or quote on a website, or trying to get something done that benefits from a certain amount of retrieval, probabilistic reasoning, or tool calling. In both cases the work involves judgement, messy information, the potential for an AI-based solution to help, and a person who still has to be accountable for the outcome.

You lead a small team that sits with the client. Each problem is a short, bounded build: design the approach, make a working version - or configure something off-the-shelf - then test it with real examples or real customers and give a straight recommendation: build, buy, or kill. The aim is to prove enough value to justify an ongoing engagement: keep finding the next problem worth solving, prove it quickly, then stop, keep going, or size what it would take to run for real.

You support how we sell and run this work: quality bar, methodology, and delivery. 

What you will do

• Support shaping and selling. Pre-sales and architecture. Turn a client problem into a bet: hypothesis, slice, team, timebox, exit criteria. Explain the approach to engineers and to a COO or professional head. Write enough of the proposal that sales is not guessing.


• Build trust through delivery. Be able to talk to the value the squad is creating, identify further opportunities; this may look like one small initial engagement that turns into RUSH being an integral partner for innovation.


• Tech-lead delivery. Small squads, high tempo. Technical direction, review of generated code, and you can change the code yourself. Distinguish a disposable proof of concept from something that could enter a release process.


• Build AI into the product. Pipelines, prompts, schemas, tool use, evals, logging, human accept/override. Design for wrong answers, missing evidence, and “the model cannot tell”.


• Release conversation. Sponsors will see a high-fidelity prototype and ask to use it on Monday. State what still has to be true: environments, scans, QA, ownership, integration, assurance.


• Playback. End of sprint: evidence, including a successful kill. Patterns the client can reuse. A sized next SoW.


• Manage the Innovation team. Assign and manage resources to deliver concurrent streams of innovation. Onboard and upskill new starters, provide inputs to L&D plans and mentor team members.

What you will bring

• Senior / Tech Lead engineering. You will have led the technical direction of a delivery, supporting design decisions, decomposition of work, review and ship. AI coding tools help us write software, but you still have a background in - and deep understanding of - core software concepts and could still write the same functionality without it (albeit with a few more searches).


• AI as part of the system. You have put models inside working products: structured outputs, tools, retrieval, grounding, evaluation, failure modes, oversight. Hands-on with LLMs in applications or workflows (agents, multimodal if valuable). Aware of hallucinations, cost, and latency as design constraints.


• Architecture under uncertainty. You can choose a functional slice, name the risks and unknowns, and change the bet when the evidence says so.


• Pre-sales. You are comfortable with ambiguous initial conversations, shaping and guiding direction, playbacks - even when they disagree with a hypothesis or recommend not doing something. You can challenge a client with evidence and experience.


• Enterprise delivery. You are aware of the eventual constraints a production-level piece of software will be under in Enterprise; confidentiality, approved tools, data that must not go in a public model, scalability.


• Prototype speed. You are able to make something to test and talk to in days, not months. You are conscious and communicative about what you cut, and you do not treat the prototype as production.

Nice to have

• Specific experience in embedded squads, innovation programmes, or timeboxed experiment funnels in large organisations.


• Examples of taking an AI-powered prototype toward something a client team can run in anger (evals in CI, observability, handover).


• New Zealand regulated or operationally critical domains (infrastructure, logistics, health, finance).