Autisleep · Yalmo
What started as a sleep monitoring hackathon project for autistic children, became a startup that grew into a full behavioural platform.
Timeline
2020–2021
Role
UX/UI Designer
Product Type
Hackathon · Seed Startup
Team
Multidisciplinary hackathon team · Clinical advisors
Tools
Figma · User interviews · Clinical research
Context
A problem hiding in plain sight
Hackautism 2.0, organised by Microsoft, is an annual hackathon focused on building tools for autistic children and their families. Between 44% and 83% of children with autism spectrum disorder experience sleep disorders. In the general population, that figure is 1–6%.
For parents and caregivers, nights are unpredictable and exhausting. Days are shaped by what happened the night before, but the connection is rarely legible. Behaviour that looks like a conduct problem is often a sleep problem. And a sleep problem is often a daytime problem wearing different clothes.
Nobody was helping families see the loop.
Autisleep
24 hours, one focused bet
Our team had 24 hours. The idea we landed on was specific: a non-contact sensor placed under the child's mattress, paired with an app for parents. Physical contact is often difficult for autistic children, so the sensor had to work without touching them. It monitored heart rate, breathing, and movement through the night, and the app surfaced what it found: what disturbed the child, how patterns were shifting, and what a better night might look like.
The sensor
Placed under the mattress. Contactless. Measures pulse, respiratory patterns, and movement through the night without waking or disturbing the child.
The app
Parents log daytime context: food, screens, activity, medication. The app correlates it with the night data and surfaces what seems to be affecting sleep.
Data over time
In the early days, all indicators appear equally weighted. As data accumulates, the system shows which factors matter most for this specific child.
First place
Hackautism 2.0. Selected for the Microsoft HUB 2021. Press coverage in Calcalist. And a question we hadn't fully answered yet.
Final design
With around two hours left after finalising the mockups, we produced the video for the competition presentation. The full system: sensor, app, and data view.
Winning team
1st place · Hackautism 2.0 by Microsoft
Selected for Microsoft HUB 2021
The Pivot
Winning opened a harder question
After the hackathon, we kept going. Conversations with families, and then with clinicians, including Dr. Lydia Gabis, head of the child development department at Tel Hashomer Hospital, pushed us toward something we had been circling without naming.
Sleep and daytime behaviour are not separate problems. Poor sleep causes behavioural difficulties the next day. Difficult days, full of noise, disruption, and stress, affect the quality of sleep that follows. The sensor could see the night. It could not see the loop.
The insight
Day
affects the night.
Night
affects the day.
Both
must be understood together.
Yalmo
Same family. Broader lens.
Yalmo was the pivot. The contactless sensor stayed. What changed was everything around it.
An AI-driven messaging bot was added to gather daytime information from parents, caregivers, and teachers at set intervals throughout the day. The system no longer asked parents to log manually at the end of a long day. It reached out when it needed to know, and kept the record quietly in the background.
At night, the sensor ran as before. In the morning, the two streams of data met. The AI correlated what had happened during the day with what the child's body had done during the night, and surfaced patterns: routines that seemed to help, environmental changes worth trying, early signals of a difficult night ahead.
Night sensor
Unchanged from Autisleep. Contactless, passive, under the mattress. Heart rate, breathing, movement.
Daytime bot
Sends messages to caregivers and teachers at set times. Gathers what happened: food, screens, activity, mood. Reduces the burden on parents.
AI correlation
Day and night data combined. Patterns emerge over time. Actionable advice for building routines that actually hold.
Multi-platform
Designed across the full system: parent app, caregiver interface, teacher-facing bot, and the data view that ties it all together.
▸ Clinical validation
Dr. Lydia Gabis, head of the child development department at Tel Hashomer Hospital, was involved as a clinical advisor from the post-hackathon stage onward. Her input shaped the evidence base behind the correlation model and helped define what meaningful data capture looks like in a clinical context.
Where I Went
From hackathon to seed, then on
I designed across the full arc: from the first 24 hours of Autisleep through the early seed stage of Yalmo, including UX, UI, flowcharts, pitch deck, and video design. As the product matured and the team grew, I moved on.
What stayed with me was the clarity of the problem, and how long it took to see it fully. The hackathon gave us a solution to one part. The research gave us the rest.
