AI Builder Challenge winner Ellen Ivanovic shares how a difficult night with her baby became Luvray, why evaluation mattered more than a high test score, and what she learned building the product solo in five days.
About Ellen Ivanovic
Ellen Ivanovic is the founder of Luvray, an AI parenting companion she built solo in five days during the AI Builder Challenge 2026. She brings over eight years of experience across product, user research, and go-to-market strategy, with a focus on responsible product design, AI evaluation, and practical tools for parents.
In just five days, you developed Luvray and won three awards. What does it do, what inspired the idea, and what makes it particularly meaningful to you?
The first thing I want anyone to know about Luvray is that it's an AI companion, not a doctor. Parents can talk to it or type to it, and it logs what they mention, like feeds, naps, a cough or a temperature, to the right child. It then uses that history to give short, warm answers, tells them clearly when it's time to call the doctor, and backs its guidance with trusted sources like the NHS, CDC, AAP and WHO rather than whatever the internet feels like saying that day.
The idea started three years ago, when my son was a baby. He had a high temperature, and my husband and I spent the whole night Googling, keeping an eye on him and trying to work out what to do. In the end we called the ER, and they told us it was fine and to keep watching him, which was honestly all we needed to hear. Even now, as he's older, we're still searching on different sites and asking different AI tools, and none of it is saved anywhere. So when he has a cough today, nothing remembers the cough he had last time. I wanted one place that keeps it all together and actually remembers your child.
It's meaningful because being a mom, and a working mom at that, is one of the most important things in my life. I never thought I'd be able to build this idea myself, and now I've built it solo, end to end, in five days. It already has its first two founding families and a waitlist, which still makes me smile every time I think about it.
With over eight years of experience in product, user research and go-to-market strategy, how did that background influence the way you built Luvray, and what has AI changed for you?
I didn't start with the code, I started with the problem. I already knew it was real because I'd lived it, and so had my friends and plenty of parents in online groups. I used a research tool I'd built myself to look at what was already out there, where those products fell short and where Luvray could fit. I thought voice would be my big differentiator, until I found that some apps already had voice agents. So I took it a step further. Luvray doesn't just listen, it remembers. It logs everything to the right child, so every answer builds on that child's history, and the whole family can share one log.
Then I thought about the whole product, not just the build. I used n8n for the backend because I'd been building with it for a while and could see every workflow clearly. Then came the front end, and the part I care about most, which is the go-to-market. You can build a great product, but if nobody trusts it, it doesn't matter. So I designed the landing page to build trust from the start, showing how Luvray uses AI responsibly and putting parents' real worries front and centre.
What AI has changed for me is that when I'm curious about an idea, I can now build it and test it myself. I have huge respect for engineers, developers and QA testers, and any product that goes to real users and scales needs that team. But being able to experiment on my own makes me a better partner to the teams I work with, and it means I never stop learning.
What was the most challenging or unexpected part of the AI Builder Challenge, and when did you realize the idea could become something bigger?
The hardest part was the evaluations, meaning checking that Luvray's answers were actually right. My tests kept coming back with really high scores, which sounds great, but anyone who works with AI knows it can be confidently wrong. So I built nearly 70 test cases based on real situations, but in five days I couldn't personally review every single one. That's what I'm working on now: going through them, deciding what a good answer looks like and what a bad one looks like, and adding more cases. My test cases come from my own experience, so once more parents start using Luvray, they'll bring situations I never thought of, and that's exactly what will make it better.
The moment I realised it could become something bigger was when a real family asked to sign up, before I'd even worked out my next steps. I was honestly taken aback. I remember thinking, oh my goodness, I actually have something here, even though part of me still couldn't quite believe it.
What did the recognition from the judges and community mean to you, and which conference moments stood out?
Honestly, I'm still taking it in. 180 builders from 34 countries took part, and to have Luvray recognised by both the judges and the community felt incredible. The community vote meant a lot to me in particular, because it came from other builders and from the people who cheered me on.
The moment that stood out most was simply winning. I built Luvray never imagining it would be chosen, let alone win three awards. The competition was tough, and I knew I was up against some highly experienced builders and I built it anyway, because I'd had the idea for almost three years and I just wanted it out there. So having my hard work and my love for building products recognised like that is what I'll remember most from this whole experience.
Straight after the showcase, I connected with several of the other builders. Everyone was so warm and open and willing to share what they'd learned, and it never felt competitive in a cold way. On showcase day it really felt like a room full of people building each other up. Each of them is working on something special, from very different backgrounds, and we're already talking about some exciting things for the future.
What did you enjoy most about the challenge and conference, and who would you encourage to join future editions?
I definitely enjoyed the building itself. It was exhausting in the very best possible way. The deadline, the pressure, and the excitement of watching it come to life made it so worth it. I'd had plenty of failed attempts at building products before, and this was one of the first times everything just came together from start to finish. Five days forced me to make decisions quickly, and somehow that made it fun instead of scary, although maybe a little scary.
I'd encourage anyone who has an idea they keep putting off to join. You don't need to be an engineer, and you don't need to have it all figured out. If you're curious, willing to learn and okay with things breaking along the way, you'll walk away with a working product and a whole lot of new people in your corner. And to the moms and dads who think there's no time for this, I built Luvray around a full-time job and a toddler, so if I can do it, you absolutely can too.