Skip to content

Site navigation

Pages of this site.

AI fitness and nutrition coaching app

FitYar

A consumer fitness app built around Borzoo, a conversational AI coach that turns a chat into a personalized training and nutrition program.

400+exercises in the library, each with video, target muscles, equipment and form guidance.

Project images

  • FitYar product screens, board 1 of 4
  • FitYar product screens, board 2 of 4
  • FitYar product screens, board 3 of 4
  • FitYar product screens, board 4 of 4

Context

What the system had to do

The product needed an AI coach that produces usable, structured training programs — conditioned on each user’s goal, fitness level, training history and available equipment — while tracking, execution and nutrition had to keep working whenever the AI layer does not.

What I built

Decisions and delivery

  • Integrated OpenAI GPT models into the product surface: designed the conversation flow, the system prompts and the structured-output handling that turn a chat session into a usable training program.
  • Built the surrounding product and data model independently of the AI layer, so core tracking and execution work without it.
  • Delivered step-by-step workout execution with set, rep and weight logging and rest timers; calorie and macro tracking against a food library; water and habit tracking; and a 400+ exercise library with instructional video, target muscles, equipment and form guidance.
  • Own backend, infrastructure and AI integration alongside one frontend engineer.

Outcomes

What it delivered

  • 400+

    exercises in the library, each with video, target muscles, equipment and form guidance.

  • Core tracking and execution run independently of the AI layer.