Hi Makers, we won a Golden Kitty Award for Health & Fitness product of the year, AMA! 😻

Aleksei Kurov
38 replies
Hi! I’m Alexey, CEO of Zenia, the product of the year in Health & Fitness Category according to Product Hunt's community 🔥 Zenia is the first AI-powered personal yoga assistant. The app uses movement recognition technology to track movements of 16 joints of human body and give gentle real-time corrections on the yoga performance. We have an extensive experience in building AI and fitness related trackers. We’ve developed over 10+ smart solutions for medical institutions and world leading athletes alike. Here we are with the team of engineers to answer all your early stage startups, AI, machine learning, movement recognition and, of course, yoga related questions. I'm answering questions alongside my great team: @dmitry_zausaev, @oleg_blokhin, @ksenya_starikova, @annayakovleva, @olga_samoilova, @bitemybyte, @cubbeewan and @ivansmolyar

Replies

Abadesi
What inspired the idea to use a phone camera to track yoga positions? 🧘‍♀️
Andreea Balasa
@aleksei_kurov Loving it! It's amazing what technology can enable us to do. I am a big fan of the kind of products that improve your lifestyle considerably (as technology should be actually used as I may add).
Aleksei Kurov
@abadesi Hi Abadesi! Thank you for your question, this is a good one. As a team we have an extensive background in working with AI and movement recognition technologies. The idea to use front camera in the yoga-related (or actually fitness niche in general) came out naturally. We’ve decided to implement it yoga, because it’s hard exercise at home only with video courses without any feedback. But during the practice you should realize if you are doing right. So Zenia helps you to practice at home with a confidence.
Aleksei Kurov
@andreea_balasa Thank you! It is inspiring us to do our product better day by day.
TL Robinson
Congratulations! That's awesome! As a newbie to the yoga space, this app will be a great help for me.
Anna Yakovleva
@tl__robinson thank you! Zenia was created having in mind the difficulties beginners are experiencing while practicing at home. Look forward to hear how your yoga practice with Zenia is going, if you have any feedback please email us at friends@zenia.co
Stefanie Lau
Congratulations on your success!
Anna Yakovleva
@stefanie_lau Thank you and congrats on the new venture! Please email us at friends@zenia.co to hear few tips from our experience.
Stefanie Lau
@zenia @annayakovleva Thank you! will do. I have downloaded it and will try tonight. I used to go to yoga 3x/week for years before I had kids, so this would be a very interesting way to re-incorporate that into my new life!
Olga Samoilova
@stefanie_lau wow, sounds great! Zenia is really very useful for home practices, especially if you don't have so much time. Have a good practice!
Stefanie Lau
@olga_samoilova Hi, sorry for the delayed response, but I have not been on Product Hunt lately! I did end up using Zenia and it was a really cool experience and the UX is very intuitive.
Olga Samoilova
@stefanie_lau Stefanie, so happy to hear your feedback - It's great that you like it :)
Stefanie Lau
My name is Stefanie and I am the Founder of Small Wins LLC. We are currently beta testing our first MVP of an app called BAMBAMsleep!, which helps moms develop healthy sleep habits in their newborns through a series of guided-challenges. I would appreciate any advice you have on marketing, specifically what marketing channels were the most successful for you? How did you get your first 1000 users? Thank you in advance.
Olga Samoilova
@stefanie_lau Hi Stephanie! Thanks for your question. We started with the classic channels: organic traffic came after the launch campaign (exclusive media publications, launch on Product Hunt); Paid traffic, of course. This is a mandatory part to get the first users’ attraction.
Olga Kazmina
Congrats! I've been following Zenia for a while and enjoy seeing your progress, guys! I've stopped doing yoga for quite a while but the idea of kinda self-coaching motivates me a lot to get back to routine :)
Anna Yakovleva
@olga_kaz3k00 Glad to hear you're getting back to yoga. If you have any questions related to practicing with Zenia, please email us at friends@zenia.co
Damla Yıldırım
Congrats Zenia Team! What a great succes :)
Olga Samoilova
@damla_yildirim1, thanks! It is only the start 🙌🏽
Pooja S
Hi there, super interesting product. While I haven't tried it yet (camera-shy android user), I'm curious on how you deal with churn. One of the most common issues with fitness and wellness is retention. Is it an important metric for you? And how do you get users to keep coming back? (Gamification, badges, push notifications, etc?)
Aleksei Kurov
@pooja_s1 Hi, thank you! Yes, churn is one of most important metrics for us. We constantly make experiments with different mechanics that can reduce churn rate. I agree with you and in Zenia we think that gamification, badges, challenges - good starting point. We believe that it will help us increase retention. Content is also very important for us. Users get bored of doing the same thing every day.
Denis Sokolov
If you want to ask any computer vision or tech related questions, you're welcome!
Victor Andrews
Congrats with your great success! Will you have an app any time soon?
Olga Samoilova
@timothy_burton Hi, Victor! Thanks for the congrats :) We already launched IOS version in December 2019. So you can try it: Zenia - Yoga & Wellness
Mikalai
Hi, it's wonderful to see your success. I one of the first people which found your Page on LinkedIn. The idea is awesome. I was inspired to learn ML and develop myself from the other side. I want to ask engineers about your tech stack. What do you use ? What can you recommend to use for movement recognition?
Denis Sokolov
@ptashytz Hi, Kolja! Thanks for the question! For model training and experiments we use pytorch and Facebook ML stack in general. To run inference on iOS we convert models via onnx2coreml and use coreML :) On Android we're experimenting with TFLite (via onnx2keras) and MediaPipe. For movement recognition we use Stacked Hourglass-like CNN with lots of tweaks and postprocessing which allows us to run it on mobile devices without losing accuracy. It's trained on our private dataset. For general purpose movement recognition I would recommend to look into OpenPose and FastPose (multiperson), Microsoft's HRNet (single person), and Google's PoseNet (both).
Mikalai
@cubbeewan Denis, thx for recommendations! If it's not a secret: are you build your dataset by yourself ( how:) ) or purchased it? And do you plan to launch your product on desktop, TV?
Denis Sokolov
@ptashytz Dataset is labelled by internal team. And yep, we're planning to support TVs.
Margarita Burkova
Hey! Congratulations 🎉 How many yoga instructors do you involve in your project and how often do you turn to them for advice?
Aleksei Kurov
@margarita_burkova Thank you! Right now we are working with 3 yoga advisers and also we have some yoga instructor helping us with testing. But we started with 1 advisor and 1 tester. Now we have a lot of content tasks so we work with our instructors every day.
Tom O'Malley
Congrats! So do they actually mail you a physical golden kitty or is it more of a virtual trophy?
Olga Samoilova
@sucantare Tom, thanks! It's both - now we have the virtual badge on our PH page, and in few days we'll receive a physical trophy - the golden kitty. But, of course, not from real gold :)