Hey HN community!
As a software engineer who took an unconventional path (bootcamp graduate in late 2021), I've spent the last 3 years in the 9-5 grind, constantly dreaming of building something meaningful. After watching my brother struggle with interview anxiety, I built Acedit: an AI interview coach.
What it does:
- Provides real-time, personalised prompts during online interviews
- Parses job listings for targeted practice questions and answers (from popular job listing websites LinkedIn, indeed, etc…)
- Offers mock interview simulations
Tech stack: NX Monorepo, Express, React, Playwright (e2e testing), Supabase (auth/DB)
What sets it apart:
- Chrome extension format (less intrusive than a PWA)
- Ability to seamlessly parse info from the users active tab (allowing for quick and easy extraction of listing information)
- Free tier with unlimited live interviews (uses gpt-4o mini)
Biggest challenges:
- Balancing helpfulness with ethical use (it's a coach, not a cheat sheet)
- Optimising AI response time for real-time use
- Building accurate "Question Detection" using NLP
I'd love your feedback on:
- Ethical considerations - how to ensure proper use?
- Performance optimisation ideas
- Potential expansions (e.g., LinkedIn lead generation, more in-depth practice modes)
Thanks HN! Excited to hear your thoughts and answer any questions.
As a software engineer who took an unconventional path (bootcamp graduate in late 2021), I've spent the last 3 years in the 9-5 grind, constantly dreaming of building something meaningful. After watching my brother struggle with interview anxiety, I built Acedit: an AI interview coach.
What it does:
- Provides real-time, personalised prompts during online interviews
- Parses job listings for targeted practice questions and answers (from popular job listing websites LinkedIn, indeed, etc…)
- Offers mock interview simulations
Tech stack: NX Monorepo, Express, React, Playwright (e2e testing), Supabase (auth/DB)
What sets it apart:
- Chrome extension format (less intrusive than a PWA)
- Ability to seamlessly parse info from the users active tab (allowing for quick and easy extraction of listing information)
- Free tier with unlimited live interviews (uses gpt-4o mini)
Biggest challenges:
- Balancing helpfulness with ethical use (it's a coach, not a cheat sheet)
- Optimising AI response time for real-time use
- Building accurate "Question Detection" using NLP
I'd love your feedback on:
- Ethical considerations - how to ensure proper use?
- Performance optimisation ideas
- Potential expansions (e.g., LinkedIn lead generation, more in-depth practice modes)
Thanks HN! Excited to hear your thoughts and answer any questions.