ASIC Engineer

The problem.

Custom silicon is the end of a road the firm has not walked yet. This listing is honest about that. What is real today is the question underneath: how much of the latency is structural, how much is ours, and at what point a fixed-function implementation becomes the only remaining move. Somebody has to answer that with authority rather than enthusiasm.

What you would do.

• Establish the argument first: where the time goes, what a fixed-function implementation would actually remove, and what it would cost to be wrong.
• Set the specification and the verification strategy before any commitment, so the decision stays reversible as long as possible.
• Work through the intermediate steps properly — software, then programmable hardware — and say when each stops being enough.
• Own physical design and sign-off if and when the programme is funded.
• Keep the correct software path alive throughout. It is what runs when the device is in fabrication.

What we would want evidence of.

• You have taken a design through to silicon and can describe what the first bring-up found.
• You can build a cost case that a sceptical reader cannot dismantle, and you have had one dismantled.
• You understand the intermediate options well enough to argue for them against your own speciality.
• You are precise about verification coverage and honest about what it does not cover.

What would make this a bad fit.

There is no funded silicon programme. This role is a poor fit if you need one to exist before you would consider a conversation. We would rather you knew that from the listing than from the second call.

This role is a poor fit if the intermediate steps bore you. You would spend most of the first year on them.