I build what's next.

Problem

• Hearing aids cost thousands, take weeks to fit, and carry a stigma most people will not accept.
• So people wait. Years, usually, and many never get past the clinic door at all.
• Meanwhile the hardware sitting in their pocket already met the spec.

Solution

• Built a mobile hearing test validated against a machine learning model trained on hundreds of thousands of audiograms and their real-world fittings.
• Wrote the resulting audiogram into Apple Health, which turns AirPods and compatible Beats into low-powered hearing aids.
• Priced it as a small monthly subscription instead of a medical device purchase, on hardware people already own.

Problem

• Growth hormone deficiency means a daily injection through childhood. Miss enough of them and you lose centimetres you never get back.
• Fear of needles is the real adherence problem, and the legacy platform treated it as a compliance issue rather than a design one.
• The ecosystem had grown into three disconnected pieces: a patient app, a clinician portal, and two connected injection devices, one Wi-Fi and one Bluetooth.

Solution

• Ran a heuristic evaluation across the whole ecosystem, then rebuilt it from scratch instead of patching around it.
• Designed personalised journeys that step a child gradually through needle fear, rather than ignoring it and hoping for adherence.
• Reconnected app, clinician portal and both devices into one loop, so paediatric endocrinologists see growth and adherence in the same place.

Problem

• Clinicians are bound to the EHR for patient data, and the EHR knows nothing about what happens between appointments.
• That gap is where adherence quietly slips, symptoms escalate and treatment decisions get made on five week old information.
• The patient-reported and biometric data existed. It just had nowhere clinical to land.

Solution

• Dawn apps collect patient-reported and biometric data, then deliver it into the clinician workflow through CLASSIFIED.
• Adaptive cards surface adherence, new symptoms and trend direction the moment the chart opens. No second login, no new interface to learn.
• Four flows behind it: Recap, Record, Signal and Sync, each tied to a real moment in the visit.

Problem

• Independent, family-run food businesses in the Nordics have no marketing budget and no route to a new audience.
• Most food content online is paid placement dressed up as a recommendation.
• People can tell, and once they can tell they stop trusting any of it.

Solution

• Built a review brand on one rule: only places worth going back to, and nothing paid.
• Shot, edited and published everything myself, focused on the family behind the counter rather than the plate.
• Kept the format short and repeatable so the audience knew exactly what they were getting.

Problem

• Independent cafes have healthy margins per drink but unpredictable volume.
• Quiet weeks and slow summers turn a good business into a stressful one.
• Nobody was solving revenue stability for them. The only tool on offer was discounting, which makes it worse.

Solution

• Roughly $50 a month for one beverage a day at any partner cafe, ordered through a Facebook messenger chatbot and sent in real time to cafe POS.
• Passed around 80% of subscription revenue straight back to the cafes as recurring monthly income.
• Onboarded independently owned cafes only, and built the brand around exactly that.

Problem

Solution