Case Studies

Real projects. Real constraints. Real results. Each case study documents a complete build — from initial problem to shipped product.

Case Study 01 MedReminder Medication Management System

MedReminder

A custom-built touchscreen medication management ecosystem. Hardware, firmware, and web dashboard — shipped as one complete product.

ESP32 Hardware IoT Embedded Systems

Delivered Q1 2026

3
Meal Windows Tracked
1
Chip Runs Everything
24/7
Always-On Operation
01

The Core Problem

Managing a daily medication routine sounds simple until you're actually doing it. Three separate meal windows. Multiple medications per window. Insulin that requires an exact unit count every single time. A missed dose isn't just an inconvenience — for someone managing Type 2 diabetes, it has real, compounding health consequences.

Phone alarms get snoozed. Phones get ignored. Generic solutions don't understand insulin dosing — a specific number of units must be injected each window, and that number needs to be recorded, not just acknowledged.

MedReminder is a purpose-built touchscreen device designed to solve exactly this. Not a phone app. Not a generic alarm. A dedicated piece of hardware that sits on a nightstand, knows what time it is, and tells you exactly what to take and when — every single day, without fail.

02

Why Hardware Instead of an App?

A phone is a distraction machine. It has notifications, social media, browser tabs, and a hundred other things competing for attention. A dedicated device has exactly one job. When you wake up in the morning and look at the nightstand, it shows you exactly what you need to see — no way to accidentally swipe away.

That single-purpose focus isn't a limitation — it's the entire product.

03

Hardware Specification

The device runs on an ESP32-2432S028R — a compact all-in-one board with a 320×240 resistive touchscreen, dual-core processor, and built-in WiFi. The entire system — UI, logic, web server, weather data, medication logging — runs on a chip smaller than a credit card.

Core Hardware
MicrocontrollerESP32-2432S028R
Display320×240 ILI9341 TFT
TouchResistive (XPT2046)
ConnectivityWiFi 802.11 b/g/n
PowerUSB-C 5V
Software Stack
FrameworkArduino / ESP-IDF
Display LibraryTFT_eSPI
Time SyncNTP over WiFi
Weather APIOpenWeatherMap
Web ServerESP32 AsyncWebServer
04

Functional Logic

Breakfast runs 5am–noon. Lunch noon–3pm. Dinner 5pm–midnight. Outside windows, a clean clock shows time, date, and live weather via OpenWeatherMap. The display is always useful, even when no medications are due.

For insulin, tapping the card opens a custom numpad — enter the exact units, confirm, and it's timestamped. Every medication also has a Skip button. Skipped doses are recorded separately from missed doses, because the distinction matters clinically.

05

The Web Dashboard

Open any browser on the home network — full dashboard, no app install, no cloud, works entirely local. Today's summary, seven-day trend graphs, and a complete timestamped history log of every dose ever recorded.

Today's Summary

Doses taken, skipped, missed + total insulin units

7-Day Trend

Visual compliance graphs across the past week

Full History

Complete timestamped log of every dose, skip, and missed window

Next Case Study
Case Study 02 Søren Smart Water Quality Monitor

Søren

A portable, AI-assisted smart water quality monitoring system. Translates raw sensor data into instant, human-readable feedback — powered by an ESP32 with a custom OLED interface and fully standalone operation.

ESP32 Hardware IoT / Sensors AI-Assisted Dev Portable

Delivered Q2 2026

Real-Time
Water Readings
2 States
Safe / Unsafe
AI-Built
Firmware & Logic
100%
Portable / Standalone
01

The Core Problem

Most people rely on visual inspection or taste to judge whether water is safe to drink. Neither method detects dissolved contaminants. In rural areas, remote regions, and even developed communities with aging infrastructure, water quality can vary significantly — and individuals typically have no practical way to assess it in real time.

Existing solutions are either lab-grade (expensive, slow, not portable) or passive (generic strips with no digital output). There was a clear gap for a device that gives immediate, understandable feedback in the field — no specialized knowledge required.

Søren was built to bridge that gap. A portable, always-ready device that measures water quality in real time and tells you — in plain language — whether the water is safe or not.

02

The Name

Søren was chosen deliberately — a human-centered name to give the project identity beyond a typical technical prototype. It reflects a shift from viewing this as just a sensor system to positioning it as a purposeful product focused on real-world impact. Good hardware deserves a name, not just a part number.

03

Hardware Specification

Every component was selected for portability, accessibility, and ease of replication. The entire device runs on a standard USB power bank — no wall outlet, no wires, no setup.

Core Hardware
MicrocontrollerESP32 Dev Board
Water SensorGravity Analog TDS (DFRobot)
Display0.96" OLED (SSD1306, I2C)
PowerUSB Power Bank (portable)
Enclosure3D Printed (in progress)
Software Stack
FrameworkArduino / ESP-IDF
Display LibraryAdafruit SSD1306
Signal ProcessingSmoothing + Noise Reduction
ConversionTemp-Compensated TDS
OutputSafe / Unsafe Classification
04

Functional Logic

The TDS sensor reads dissolved solids in real time. Raw ADC values are processed through a smoothing algorithm to reduce noise, then converted to parts-per-million (ppm) using a temperature-compensated formula. This conversion is critical — raw voltage readings are meaningless without it.

The ppm value is classified against established safe drinking water thresholds. Instead of showing a technical number, the OLED outputs a direct, human-readable verdict.

Safe Reading
✓ Water Safe

TDS within acceptable threshold — tap water typically 100–200 ppm

Unsafe Reading
✕ Water Unsafe

TDS above safe threshold — salt water tests at 1,000+ ppm, clearly flagged

05

AI-Assisted Engineering

Søren demonstrates what modern AI-assisted development looks like in practice. This wasn't AI generating a finished product — it was AI acting as a collaborative engineering tool that compressed the development timeline dramatically.

The conductivity-to-TDS conversion and temperature compensation logic — normally requiring days of research and debugging — was completed within hours using AI-guided code generation. The same approach was applied to ADC inconsistencies, signal smoothing, and overall system architecture.

Firmware Dev

Sensor logic, ADC calibration, and sampling routines built with AI assistance

Conversion Logic

TDS formula and temperature compensation — implemented in hours, not days

Debugging

ESP32 ADC inconsistencies and signal noise diagnosed and resolved rapidly

06

Real-World Relevance

Access to safe drinking water remains a critical issue in many parts of the world — including rural and remote communities in Canada. Søren addresses this with a portable solution that delivers immediate feedback without requiring specialized knowledge or equipment.

Ideal Use Cases

  • Rural and remote communities
  • Northern Canada and Indigenous communities
  • Outdoor and recreational water sources
  • Educational and awareness programs
  • Emergency preparedness kits

Testing Results

  • Tap water: 100–200 ppm (flagged safe)
  • Salt water: 1,000+ ppm (flagged unsafe)
  • Real-time updates with stable readings
  • Reliable differentiation across water types
  • 3D-printed enclosure in progress

More Projects Coming

New case studies added as projects are completed.

Hardware / Firmware

Undisclosed Project

Industrial sensor integration with custom dashboard and OTA updates.

Coming Soon
Web App / SaaS

Undisclosed Project

Full-stack SaaS MVP from concept to first paying customer.

Coming Soon
IoT / Data

Undisclosed Project

Real-time environmental monitoring with alerting and analytics.

Coming Soon

Want us to build yours?

The first 30 minutes are free. Bring your idea — we'll tell you exactly what it would take.

Book a Free Session →