Smart Rewards for
Better Training

01 / Problem

A Community Failing its Trainers

Many dog trainers depend on remote reward-based training tools, yet existing options are outdated, costly, and disappearing from the market. The go-to device has been the same for over a decade with zero meaningful innovation — and it's increasingly hard to source outside the US.

These limitations undermine effective, science-based training in real-world environments and are creating a growing gap in the professional dog-training market.

The real-world context every hardware decision was designed around

02 / Research

What We Heard

We surveyed active dog sport trainers and conducted competitive analysis across Treat & Train, Furbo, and other generic alternatives. The frustration was immediate and unanimous.

"I would pay a lot for a device that didn't jam constantly. I used to use the MM a lot more than I currently do — and it's all due to the jamming."

— Active Dog Sport Trainer, Survey Respondent

Evidence collected from a survey of active trainers confirmed that 76% named jamming as their #1 issue. Limited treat compatibility, no app integration, and poor stability rounded out the top frustrations.

Form factor and scale — designed to sit comfortably in a training environment

03 / Personas

Who We Designed For

Two distinct user types emerged from research, with different needs — but the same core frustration.

04 / Research → Design

Every Feature Earned Its Place

Research findings mapped directly to hardware decisions. Nothing was added without a trainer pain point behind it.

05 / My Role

Learning Hardware From Scratch

Starting point: nearly zero hardware experience. I had used an Arduino only once before, with no electronics background — and we needed a BLE servo dispenser built quickly and reliably.

I chose a servo motor (DS3218) over a stepper because it detects and self-corrects positional errors. I resolved an ArduinoBLE + Servo library timer conflict through targeted research, implemented a sweep-and-return cycle (0°→180°→0°) with buzzer feedback, and connected the device via BLE GATT characteristic — tested live with nRF Connect. All wiring was managed with Dupont wire connections, no soldering iron.

I used Claude as an on-demand tutor — for servo behaviour, BLE GATT structure, and timer conflicts — alongside consulting domain experts throughout the process.

Arduino BLE / GATT Servo Motors UX Research Figma AI-Assisted Learning

The build process — from schematic to wired prototype

#include <ArduinoBLE.h>
#include <Servo.h>

// Sweep-and-return: dispenses one treat per call
void sweepAndReturn() {
  Serial.println("Sweep start");
  beep(1, 80);
  myServo.write(180);
  delay(600);   // time to reach 180°
  beep(1, 80);
  myServo.write(0);
  delay(600);   // return to base
  beep(2, 60);
  lastAngle = 0;
  Serial.println("Sweep done");
}

void loop() {
  BLEDevice central = BLE.central();
  if (central) {
    while (central.connected()) {
      if (servoChar.written()) {
        String input = servoChar.value();
        input.trim();
        if (input == "sweep") sweepAndReturn();
      }
    }
  }
}

Final Arduino BLE code — servo + buzzer feedback loop

Presenting CuePo — the final pitch

06 / Prototype

Cardboard to Working Device

We moved fast — from a rough cardboard prototype proving the mechanism, through foam and 3D modeling, to a fully wired Arduino prototype with BLE communication. Here it is actually working.

Working prototype — BLE command triggering the servo-driven agitator via nRF Connect

Cardboard v1 — proving the mechanism

Foam prototype — form and ergonomics

Components — servo, BLE module, buzzer

Team inspection — catching issues early

Wiring diagram — Arduino Nano 33 BLE · servo (DS3218) · buzzer · Dupont connections

07 / App Prototype

The Companion App

Three core flows designed in Figma — onboarding, live training session, and the dashboard. The app turns raw button presses into meaningful training data.

Onboarding — pairing the device and setting up your first session

Training session — live reward delivery, rep counter, and jackpot mode

Dashboard — session history, success rates, and progress over time

09 / Reflection

What I Learned, What's Next