Is it okay to include an LLM project in a new grad portfolio?
I’m preparing for entry-level backend roles, and I built something like an internal document search chatbot using the GPT API. But I’ve heard lately that including AI projects can actually work against you, so I’m not sure.
From an interviewer’s perspective, is a project that just calls an API not very impressive?
4 answers
Well, the claim that "AI projects these days are a net negative" feels a bit outdated. That sentiment comes from the flood of portfolios that just copy the same tutorials—properly built ones are still a net positive. I'm curious where that claim came from.
It's true that just making API calls can be done anywhere, so it can look like there's no differentiator. But I'm not saying to leave it out entirely—if you show the thinking behind things like your retrieval pipeline, chunking strategy, and how you handled prompt failure cases, that can still make a strong impression. In fact, these days few entry-level backend candidates have worked with LLMs, so the subject matter itself is good.
Interviewers care more about why you chose a given design than about the final result. If you added RAG, they’ll ask why you used a vector DB, whether you considered hybrid search, and how you managed token costs—and it’s true that failing to answer can backfire. Only include what you can explain yourself.