Backend Go + python or only python for backend with LLM ?
Hello I want to build a project and the main feature needs LLM to notice pattern from different type of object (image, test, audio).
the workflow :
- User forward object
- Object goes in a S3
- Object are processed to facilitate the work of the LLM
- AI works to get pattern
- User sees results on frontend
The first three steps can be done in both language, the fourth better ecosystem in python (I guess) and the last one in ts but it's not the problem.
Basically all the project I did were in go, but I learned python and it's not difficult to have hands on. My goal is not to build skills on a specific language, it's to have a working prototype, but I don't want to vibecode it without understanding the code.
I planned to do it in go + python because go is really cool, I like it, I have media to process, queue and jobs, form what I see on reddit it's a better backend language but i'm afraid to lose time to wire the two languages, because harder local dev, harder debugging, more CI/CD work, slower iteration, two types to write and observability (okay this is probably gonna be vibecoded).
In the other hand doing it only with python will slow me down on the backend part, I know how to build api, authentication, structure a project, deploying the project with go, and in python I did a little bit of scripting and exercises online and it's the best to manage llm from what I see (in both case I have to learn how to use llm btw).
so or Go API get request + Queue -> Python API receive the task and control LLM -> LLM -> next
or Python API get request + Queue + control LLM -> LLM -> next
(Maybe I'm wrong with this model and Go API can control the LLM)
So do you think I should start with Python + Go or just Python ?