The freeCodeCamp Podcast

#198 When NOT to use AI in your hackathon project with MLH winners Cindy Cui and Alison Co

Nov 21, 2025
Guest
Cindy Cui
Guest
Alison Co
Alison Co and Cindy Cui, talented software engineering students and hackathon winners, discuss their project BrailleLearn, which helps visually impaired individuals learn Braille. They share insights on the challenges of low literacy rates and the importance of accessible tools. The duo dives into the role of AI in their development process, stressing its limitations while emphasizing hands-on learning. They also reflect on securing internships, the value of teamwork, and their experiences at GitHub Universe, revealing how collaboration can spark innovative solutions.
Ask episode
AI Snips
Chapters
Transcript
Episode notes
ANECDOTE

MacGyvered Braille Prototype

  • The team built a tactile Braille prototype using dollar-store parts, sewing pins, duct tape and late-night MacGyvering.
  • They iterated through mechanical failures until a pin moved in a straight linear motion around 5 AM, which proved the design worked.
ADVICE

Prefer Local Rules For Real-Time Tasks

  • Avoid adding LLM steps that introduce latency when real-time responsiveness matters; prefer deterministic local rules.
  • Use custom speech grammars for letter recognition instead of slow external LLM calls in live demos.
ANECDOTE

From Finalists To GitHub Universe

  • The team learned they were finalists via a surprise phone call and experienced repeated 'pinch me' moments after winning.
  • GitHub invited them to present at Universe, where they demoed the project and met accessibility leaders and Satya Nadella.
Get the Snipd Podcast app to discover more snips from this episode
Get the app