Design a Mobile Workout / Fitness Coaching App

By /
⏱ 3 min read

“Design a fitness coaching app” is the structured-workout cousin of the Strava-style activity-tracking question. Peloton App, Nike Training Club, Apple Fitness+, Beachbody (BODi), Future, and Centr are the references. The interview tests video delivery, sensor integration, plan progression, and the unique UX of guided exercise.

Clarify scope

  • On-demand workouts only or live classes?
  • Personal trainer / 1:1 coaching?
  • Equipment-based (Peloton bike) or bodyweight?
  • HealthKit / Google Fit integration?
  • Apple Watch independence?

Workout data model

{
  workout: {
    id, title, instructor, duration,
    discipline: "strength|cardio|yoga|...",
    difficulty,
    exercises: [
      { id, name, sets, reps, durationSec,
        startMs, endMs,  // in the source video
        equipmentNeeded }
    ],
    videoAssetId
  }
}

Video delivery

  • HLS streaming with ABR
  • Pre-cached intro/outro to start fast
  • Multiple bitrates; default 720p on cellular
  • Captions for instructor cues
  • Seekable timeline with exercise markers

Plan progression

  • Multi-week plans (e.g., “8-week strength”)
  • Each day has a workout assigned
  • User progress tracked; can skip or repeat days
  • Adaptive: if user falls behind, plan adjusts (push deadlines or drop optional days)

Live class infrastructure

  • RTMP ingest from instructor; HLS-LL for low-latency delivery
  • Live leaderboard with avatar of participants
  • Real-time metrics from connected equipment (Peloton bike: cadence, resistance, output)
  • Chat / shoutouts (Peloton-style)

Sensor integration

  • iOS: HealthKit for HR, calorie, workout sample writes
  • watchOS app: heart rate during workout
  • Bluetooth HRM (Wahoo, Polar) for users without watch
  • Smart equipment: Peloton bike, Tonal — proprietary protocols
  • Step / distance from accelerometer for cardio

Apple Watch app

  • Independent workout app on watch (no phone required)
  • Activity rings update during workout
  • Mid-workout metrics (HR, time elapsed, current exercise)
  • Haptic cues for set transitions
  • Audio coach if AirPods connected

HealthKit integration

  • Read: HR, weight, height for personalization
  • Write: workout samples (auto-counted in Activity rings, contributes to Move/Exercise)
  • Permissions per type — granular
  • HealthKit privacy: data stays in HealthKit; export to your servers only with explicit consent

Progress tracking

  • Workouts completed (this week, all-time)
  • Total minutes, calories, average HR
  • Streaks
  • Per-exercise PR tracking (max weight, longest hold)
  • Weekly and monthly summaries

Personalization

  • Onboarding survey: goal, level, equipment available, time budget
  • Recommended workouts based on history
  • “Pick up where you left off” front-and-center
  • ML for next-best-workout based on completion patterns

Offline downloads

  • Download for travel
  • DRM if licensed content
  • Auto-download next workout in plan on Wi-Fi
  • Storage cap with auto-cleanup of completed workouts

Notifications

  • Daily reminder at user-set time
  • Streak protection (“Today is your last chance”)
  • Live class starting soon
  • Personal milestone (10 workouts, 1000 minutes)
  • Per-category opt-out

Audio cues

  • Voice coach overlay on the workout video
  • Volume mixing: video lower, coach louder
  • “Halfway there”, “30 seconds left”, “good form”
  • Music: licensed library or user’s own (Spotify/Apple Music integration)

Form / pose detection (advanced)

  • On-device pose estimation (Apple Vision pose, ML Kit pose)
  • Compare user’s motion to reference; provide feedback
  • Used by Tonal, Mirror, Peloton Guide for at-home form correction
  • Heavy CPU; run only when explicitly enabled

Social features

  • High-five / cheer on friends’ workouts
  • Compete on leaderboards (Peloton)
  • Share completed workout to social
  • Privacy: per-friend visibility

Subscription model

  • Most apps subscription-based ($10–$40/month)
  • Family sharing on iOS for paid plans
  • Annual discount
  • Free trial with paywall after first workout or first week

Performance considerations

  • Battery: video + sensors are heavy; throttle non-critical features
  • Memory: release video buffers behind playback head
  • Audio routing: AirPods vs speaker vs car; handle interruptions
  • Background: lock screen controls work; user can lock the phone mid-workout and audio continues

What separates senior from staff

Senior candidates draw the workout data model and video playback. Staff candidates discuss HealthKit integration, Watch app architecture, and live class infrastructure. Principal candidates address pose detection, the personalization ML loop, and the equipment-integration story (proprietary sensors).

Frequently Asked Questions

Should the workout be a video or a sequence of timer screens?

Both. Video for follow-along; timer screens for self-paced strength. Many apps use timer-only for some disciplines and video for others.

How do I handle the watch as the primary device?

For Apple Watch independent workouts, watch app starts the workout; phone follows when present. Sync state via WatchConnectivity. Strava and Workout app are the references.

What about AI coaching?

Increasingly common: LLM-generated personalized coaching messages, voice cues that adapt to your performance. Out of core scope but a 2025+ differentiator.

Scroll to Top