US Remote Solidity Take-Home Assignment: Gas Optimization vs Clean Code — Do Interviewers Actually Grade SSTORE/SLOAD and Foundry Tests?

I’m interviewing for US remote Solidity roles and keep getting a Solidity take-home assignment instead of a live coding round. The part I’m stuck on isn’t “can I finish it” — it’s what they actually grade.

Typical prompt: implement a small contract (ERC20-ish / vault-ish), add a couple of edge cases, and ship it with Foundry tests. I can make it “clean” (natspec, readable modifiers, clear revert reasons, consistent naming, good test structure). But I can also go down the rabbit hole of gas optimization interview style changes: caching, tight storage packing, minimizing external calls, and micro-optimizing SSTORE/SLOAD patterns.

Here’s the dilemma: in a take-home, do US interviewers reward “smart” gas wins, or do they penalize anything that hurts readability/maintainability? Like if I use assembly for one hot path, or I refactor to reduce storage writes but the code becomes harder to review.

If you’ve been on the hiring side: what are the usual grading buckets?

• correctness + security basics

• clarity + architecture

• test quality (Foundry tests)

• gas profiling / measurable improvements

• “production thinking” (docs, tradeoffs, assumptions)

Also: is it worth including a short write-up (what I optimized, what I didn’t, why), or does that feel like overkill?