We are excited to announce that we have started a self-training program for our engineers called “HACARUS Internal Kaggle Challenge”
In this program, every engineer is allowed to spend up to 20% of working hours for a Kaggle competition and this is, so to say, Hacarus version of 20% rule.
Needless to say, Kaggle is the world’s largest machine learning competition platform and we can find many competitions related to our focused business areas like medical and industrial fields. Participation to such competitions will be good chance for us to deepen our knowledge and acquire new, useful insight to fuel product development. Thus we are carefully choosing competitions for our program.
Since every engineer has his/her own technical interest, this training program is not mandatory but voluntary. We plan to have this program 3 or 4 times in a year and our engineers can choose which ones to tackle with. We have 7 engineers for our first batch which is targeted at image analysis competition for diabetic retinopathy.
What’s special about Hacarus program is that it is not only for data scientists, but for all engineers at Hacarus – such as our software development engineers, who develop applications to wrap machine learning model and our hardware engineers, who optimize machine learning algorithm for hardware. Their technical expertise is not machine learning – to remedy this, Hacarus engineering leadership believes that participating engineers should know:
- The Basics of machine learning
- The Entire workflow of building a machine learning model
- The Typical use case where machine learning methods can be used
This is to ensure smooth communication and mutual understandings across teams, which really matters for good product development.
To fill skill gap or to let members, who haven’t worked closely so far, work together, we strongly encourage to form team with at least 2 people.
“To be honest, I wanted to allow everyone to try this program but educational cost seems too high if he/she doesn’t know how to code and ultimately has to give up. However, as a member of AI startup, everyone at HACARUS is supposed to play with Jupyter Notebook even if he/she works as sales, HR or admin or else. I’ll think different types of training program for it, anyway” said Takashi Someda, CTO.
When the first batch is over, we’ll share updates via our social media and blog – including insights and lessons learned throughout the competition – so stay tuned!
We keep investing in building novel AI technology and our people here at Hacarus – if you have interested in working with us and develop your career as data scientist, please visit our career page!
