The Role

We are looking for a technically deep and commercially astute Solutions Architect to own the technical heart of our customer engagements as we bring the world's first optical compute inference platform to market. You will be the person who turns a customer's interest into a credible technical opportunity.

You will lead deep technical discovery with the most demanding customers in the industry – hyperscalers, neoclouds, frontier AI labs, and national laboratories – and translate what they need into a Lumai-based solution. You will work hands-on with our benchmark team to model and prove performance against real customer workloads, and you will author the technical proposals and responses that allow a customer to make the argument for Lumai inside their own organisation.

This is an early-stage role where you will help write the playbook. You will be as comfortable in a benchmarking discussion with a customer's ML and performance engineers as you are presenting a solution to their technical leadership, and you will act as a vital channel back into our own product and engineering teams, ensuring the voice of the customer shapes what we build next. In the near term, much of the work will be focused on pre-hardware performance modelling and simulation.

What You Will Do

Partner with Lumai’s VP Business Development, owning the technical relationship with each customer

Conduct deep technical discovery with customer ML, inference and infrastructure engineering teams – their models, serving stacks and performance targets

Translate customer requirements into a Lumai-based solution, mapping needs against both near-term and roadmap solutions

With business development, define the shape of an engagement – benchmark, proof of concept, reference architecture, or statement of work

Work hands-on with Lumai's benchmark team to scope, run, and interpret benchmarks against GPU and other baselines, and own the presentation of results

Be the voice of the customer in refining Lumai’s customer-facing simulation framework, and support customers in running their own workloads on it

Build performance and TCO models – throughput, latency, cost and energy per token – showing how Lumai compares with incumbent solutions on customer workloads

Produce the technical case for Lumai – proposals, solution documentation, and responses to RFIs, RFPs, and tenders

Design and deliver demos and technical workshops for customer technical teams

Work with engineering to define how Lumai integrates with customer software stacks – serving frameworks, runtimes and orchestration

Feed customer requirements, evaluation feedback, and roadmap reactions back into product and engineering, informing the Lumai roadmap

What We Are Looking For

Must-have

At least five years’ experience in a solutions architecture, technical pre-sales, or senior field engineering role within AI infrastructure, high-performance compute, semiconductors, or a comparable deep-tech hardware domain

Strong working knowledge of LLM inference at scale – model architectures, serving frameworks such as vLLM, TensorRT-LLM or SGLang, batching and parallelism, and the trade-offs between throughput, latency and cost

Hands-on experience benchmarking and analysing the performance of AI workloads, including performance modelling, with the credibility to defend results to customer ML and infrastructure engineers

Demonstrated ability to lead technical discovery with sophisticated customers and translate requirements into a solution

Comfortable operating in a fast-moving, sometimes ambiguous, pre-product environment

Desirable

Direct experience selling into, or working within, hyperscalers, neoclouds, frontier AI labs, or national laboratories

Awareness of the AI accelerator competitive landscape

Background or genuine curiosity in novel compute – photonics, optical computing, or other post-silicon architectures

Experience with accelerator software stacks – compilers, kernels and runtimes – or bringing a new accelerator into an existing ML ecosystem