After four years since first initially starting work on ScentSee, the perfume recommendation assistant, I have decided to pull the plug on it at the end of this month (September, 2019), due to limited commercial interest on the local and regional markets.
I am grateful to the partners and clients we have worked with, especially to Baneasa Shopping City, for their trust, commitment and for believing in home-grown innovation. We definitely thank them for being, at the time of our collaboration, the client who made us cross the break even line.
The path of transforming a sensory, olfactory experience into an analytical, mathematical model based on note types, placements and similarities turned out to be my shy first step towards discovering latent space embedding, computer vision and visual machine learning.
It is for this reason that I want to express my gratitude by making the source code for the ScentSee backend, API services and front-end UI open source on Github:
Please note that, as our team is currently focusing on our visual machine learning projects, Envisage.ai and Knosis.ai, our bandwidth for providing any support on Scentsee is at most limited. The repositories are provided “as is”.
The live deployment of ScentSee is also live until the 25th of September 2019, so if you want to try it out without having to install the open source version, now is your chance 🙂
We strongly believe that failure is the heat of battle that forges an entrepreneur. In this spirit, we hope that our pre-Docker experience with ScentSee and my (really badly-written) code shall provide some insight into the pains, complexities and mistakes inherent to bootstrapping an MVP.
Furthermore, we hope that by our small gesture we encourage other entrepreneurs to also be open about failure, since that is a sure way to turn it into an invaluable set of learning for the community.
In the world I want to live in, when startups and companies die, they go to open source heaven.