Show HN: Open-Source AI Interview Trainer(github.com)
github.com
Show HN: Open-Source AI Interview Trainer
https://github.com/dcrebbin/ai-interview-trainer-frontend
1 comments
Backstory:
I originally made a version of it at a previous job awhile ago as an innovation project but didn't get too much traction so have recreated a better version using a completely different stack.
My newer version was also originally tailored just for people looking to get a job in the Australian Public Service but have now pivoted it to big tech (as I have an interview coming up )
There are currently 2 interview modes, Problem Time Trial and Algo Kata but more to come in the future
Problem Time Trial was using questions scrapped from leetcode, thinking about it, probably a big no no so they're removed.
I originally made a version of it at a previous job awhile ago as an innovation project but didn't get too much traction so have recreated a better version using a completely different stack.
My newer version was also originally tailored just for people looking to get a job in the Australian Public Service but have now pivoted it to big tech (as I have an interview coming up )
There are currently 2 interview modes, Problem Time Trial and Algo Kata but more to come in the future
Problem Time Trial was using questions scrapped from leetcode, thinking about it, probably a big no no so they're removed.
would be cool if you support ollama in addition to openai
Local LLMs are defo the way forward for governments and certain private companies, that was a huge blocker with making these types of products.
Haven't used ollama but I've used GPT4ALL and that has the ability to turn on a simple rest api that mimics the standard openai completions endpoint, so would be super easy!
Haven't used ollama but I've used GPT4ALL and that has the ability to turn on a simple rest api that mimics the standard openai completions endpoint, so would be super easy!
Demo: https://www.youtube.com/watch?v=ef2ivitjiBU
Allows for a relatively seamless big tech coding interview experience.
Frontend: Qwik
Backend: Go Fiber
Database: MySQL (GORM)
AI Providers: (Vertex only has hacky support)
Text Generation: OpenAI, Vertex
Text to Speech: OpenAi, ElevenLabs, Unreal Speech, Vertex
Speech to Text: OpenAi (Whisper), Vertex
The code is definitely not perfect as this is my first project using Go & Fiber and probably (definitely) has some security flaws in there, but it's all an iterative process.
The real project is currently being tested by a couple of mates so nothing major (Yet question mark?)
Big fan of positive and negative feedback so lmk what you think!