One engineer. Complete pipeline. Working hardware.
You know exactly what needs to be built. You understand the domain constraints, the regulatory requirements, what success looks like in the field. What you don't have is time to become a hardware expert just to shepherd a project through five different specialists.
Most hardware development feels like throwing money into a black box. You explain your requirements to one engineer, get gerber files from another, and somehow you're supposed to navigate the factory yourself. By the time you realize something's wrong, you've already spent the budget.
Let's talk about your project
hello@ardwell.co.zaYou get one engineer who owns everything from the first schematic to the boards arriving at your door. No coordination overhead. No finger-pointing when something breaks.
I design circuit boards that can actually be manufactured at scale. High-speed routing, power regulation, communications interfaces —all designed within assembly constraints and long term component availability in mind.
You get: 3D renders before ordering, transparent BOM costs, boards that match your enclosure on the first spin.
I write the code that makes your hardware reliable. ESP32, Nordic nRF52, and bare-metal C for power-critical applications. State machines that handle edge cases. Sensor fusion that produces clean data, not noise.
You get: Firmware that survives field deployment, not just bench testing. Solid documentation on how to interface and interact with your product
Hardware is useless if you can't see what it's doing. I offer the full data pipeline — MQTT ingestion, PostgreSQL storage, Grafana or custom dashboards so your team can monitor devices in real-time.
You get: Live data you can trust. Alerts when something fails. Historical trends without manual CSV exports.
Hardware challenges I've solved across medical, academic, industrial, and commercial applications. Your domain is different — the process is the same.
A Professor at Stellenbosch needed an automated IR spectroscopy measurement system for bioreactor research. Their existing prototype was abandoned by a student and didn't have enough channels for all the required sensors.
Result: A custom 10-channel multiplexed system. Live data logging implemented to Google Sheets so no manual data extraction was needed.
An electric golf cart battery manufacturer needed an interface and communication board that could survive field deployments, a wide input voltage range, reverse polarity, and host a IoT microcontroller.
Result: 65V-rated hardware with relays and a configurable CAN, RS485, or TTL comms system. Repeat order a year later for more boards.
Groote Schuur's radiography department was spending 12+ hours per month on Excel-based staff rotas. Manual shift swaps, overtime miscalculations, fairness complaints.
Result: Constraint-based optimization with Python OR-Tools. Schedules 40+ staff in seconds. Fairness analytics built-in.
An Escape Room needed 15+ interactive sensors across 3 rooms that could run continuously for 8 hours a day on-site. The project included a local dashboard for game master control and overrides, and multiple fallback and failure prevention mechanisms
Result: Distributed ESP32 mesh network. RFID triggers, magnetic actuators, real-time coordination.
Hardware development shouldn't feel like a gamble. Here's the process I use to keep you in control while I handle the technical complexity.
You bring deep expertise in your domain—whether that's medical devices, industrial control, scientific instrumentation, or something entirely different. I bring the hardware engineering.
The problem with most engineering consultants is they treat your spec sheet like gospel, disappear for 8 weeks, and then deliver something that technically meets requirements but completely misses the point.
"I'm terrified of spending my budget on hardware that doesn't actually solve the problem."
I get it. That's why transparency isn't optional—it's my entire business model.
We start by talking about what you're actually trying to measure or control. What are the real-world constraints? What failure modes keep you up at night? I don't care about your preferred microcontroller—I care about whether this needs to work at -20°C or survive vibration testing.
You get regular updates as I design: 3D renders of the boards, BOM breakdowns in plain English, physical dimension checks before anything goes to the factory. If I'm making a tradeoff between cost and performance, you'll know about it before the order is placed.
I don't hand you gerber files and wish you luck. I manage the component sourcing, talk directly to the assembly house, and deal with the inevitable "this part is out of stock" emails. You get a working product, not homework.
Manufactured boards arrive at my bench for validation. I flash the firmware, run your test scenarios, document any quirks, and then ship you a complete system—not a pile of components that need debugging.