As a web developer who's been laid off and is now on the job hunt, I'm struggling with take-home assignments. It seems like the expectations have changed dramatically. In the past, companies focused on clean code and logic, but now I face projects where the requirement is to indicate what parts of the work were AI-generated versus manual coding. I misunderstood this to mean a 50/50 split, but I was rejected after submitting a project that didn't meet expectations, while others used AI to build out significantly more features in a fraction of the time. In a recent interview with a well-respected company, my focus on code quality didn't match what they were looking for. I'm confused about how to balance quantity versus quality in this new landscape and whether security and maintainability are being overlooked. What should be my approach moving forward?
9 Answers
The feedback you've received suggests that you should leverage AI more aggressively for these tasks. Focus on creating a fully functional product instead of getting hung up on how clean the code looks. In this environment, delivering more features is being rewarded over polish.
In interviews, I focus more on the end results than how the applicant got there. If the final product is functional, clean, and well-documented, I'm satisfied. This shows that they have solid fundamentals and can smartly use AI, combining reasoning with output.
Expect to be cleaning up AI-generated code for quite a while. Companies will still need people who can manage and improve it, so be ready to adapt.
If you're not learning how to leverage AI tools effectively, you might fall behind in this new environment.
Always ask the company about their preferences regarding AI involvement during assignments. Different companies have different views, so being upfront can help align your approach with theirs.
Consider it like an open-book test. Be transparent about your AI use, maintain a simple log of your decisions, and be prepared to discuss any trade-offs during the review. Prioritize making your work easy to evaluate, with small updates, tests for crucial functionality, and comprehensive README files.
The request for you to label AI work is a good sign—they want to see understanding rather than just rapid output. Use AI to handle the repetitive tasks while keeping the core logic and architectural decisions distinctly yours. That's the key they are looking for.
My advice would be to clarify with the employer how they view the use of AI in projects. It's better to ask directly rather than guess and risk wasting time. Knowing their stance upfront can save you a lot of trouble later on.
The best approach is to treat AI like a junior developer that you're mentoring. You should write the overall architecture and key decisions yourself, using AI to help with the less critical parts. For instance, you might say, 'I designed the API and data models myself and used AI to help with CRUD endpoints, while I wrote the authentication middleware manually to ensure it meets my logic requirements.' Being explicit about your role showcases your understanding and initiative.
It sounds like some companies have adopted a startup mindset, where your worth as a developer is judged by how many lines of code you produce daily instead of the quality of that code.
I don't think that's true. If you can create tools that utilize AI to automatically clean up code, you'll be ahead of the game.