I won't repeat what others have said about advances in natural language processing since 2017, but it's true that it's a solved problem if your problem isn't "perfect search" but a more realistic "excellent search".

> Use existing technologies first: As in most engineering problems, don’t reinvent the wheel yourself. When possible, use existing services or open source tools. If an existing SaaS (such as Algolia or managed Elasticsearch) fits your constraints and you can afford to pay for it, use it.

I work at an AI search company (Relevance AI) and even we see that InstantSearch.js is all some people need in terms of UI. We created a version of it that uses our NLP-backend but is still the same Algolia components on the frontend: https://www.npmjs.com/package/@relevanceai/instant-relevance

The reason was because those components work. Think these days you'd need to ask a lot of questions before completely rolling your own UI or NLP handling.