Hi, my name is Takashi, I am CTO at Hacarus.
As Hacarus is currently focusing on hiring new team members, we meet many new people and everytime introduce ourselves to them. At a point we thought it could be helpful to write our self-introduction down, as a reference for people we spoke to, and as a way to share and introduce ourselves to more.
This article is meant for everyone who is interested in working in Kyoto, Japan, or wants to know more about machine learning or sparse modeling at Hacarus. Read on to learn about our current development environment at Hacarus and the topics we are working on.
Hacarus’s Business
Company Information
Hacarus is a startup company, based in Kyoto and focused on machine learning. When Hacarus was founded in 2014, we started as a healthcare application company providing IoT hardware. At that time, we tried to cloud-fund our hardware product on Makuake, the Japanese version of Kickstarter.
From the beginning, we developed our product using machine learning technology. Very soon, through meetings with clients in the medical and industrial fields, we noticed that we can utilize our technology, skills and strengths to solve the problems they have.
Currently, we develop AI solutions and provide data analysis services using sparse modeling, with a focus on medical and industrial use cases. Medical use cases include disease diagnosis or treatment support based on human or animal patient health parameters or images (X-ray, CT scans). For industrial use we provide lightweight AI solutions that can be embedded into edge computers and industrial equipment (through FPGA) and video analytics solutions for example for defect detection in industry production.
Financing
Hacarus’ received funding not only from VCs, but also from companies operating in medical and industrial fields. Through these fundings, we have strong business partner relationships that are basis for business development. This is particularly helpful with regards to getting access to data that other companies cannot easily obtain from medical or drug development environments. (Needless to say, extreme caution is required as well). Therefore, these collaborative relationships accelerate our solution development.
See also below link.
Development Environment
Below we list the main tools our teams use on a day-to-day basis:
Software Engineering
- Docker
- Git
- Jenkins
- Python 3.x
- Swift ( for iOS )
- Kotlin / Java ( for Android )
- Vue.js / ReactJS ( for frontend )
Originally we used ReactJS for front-end development, but recently we switched to use Vue.js for light projects. The common language between software engineers and data scientists is Python. Several of our team members have presented at PyCon conferences before.
Data Science
- Python 3.x
- pandas / NumPy / SciPy / scikit-learn / scikit-image / seaborn / opencv / spm-image ( Hacarus owned OSS )
- Jupyter
Our Data Scientists mainly do data analysis using the combination of Python and Jupyter. We combine other open source solutions according to the nature of projects. At the same time we continue to work on our own spm-image library, that is available as opens source.
Embedded Engineering
- C / C++
- Xilinx SDx
- integrated development environment of FPGA provided by Xilinx
- OpenWRT Linux / Alpine Linux / Ubuntu ( For embedded enviroment )
In environments that support Linux, we use Python on Docker. For FPGA system, we write the code using C / C ++, doing high-level synthesis through SDx.
Infrastructure
- Ansible
- Fabric
- Terraform
- Serverspec
- AWS
- ELB / EC2 / Cloudfront / S3 / Route53 / Cloudwatch / SES / ElastiCache / MySQL RDS / Lambda / SageMaker and so on...
Communication Tools
- Backlog
- Slack
- Google Meet / Zoom
Developer Teams
Data Scientist Team
Our CDO (Chief Data Officer) leads the Data Science team members, former Kyoto University researchers, and part-time data scientists.
They are responsible for model construction based on our clients' needs and and executing machine learning with the obtained data.
Master's and undergraduates of the information technology at Kyoto University and master's degree and undergraduate degree students from Tohoku University are supporting as interns. Some of them work remotely.
Application Developers Team
The VPoP (Vice President of Product) in Kyoto leads our team members in the Philippine branch. Currently the team members in the Philippine mostly work remotely from home.
The application development team develops our SaaS, web and mobile applications and operates our server infrastructure.
An undergraduate of information technology at Kyoto University joined the project of smartphone application development as an intern.
In addition, we work closely with academic advisers from famous Japanese universities:
We are hiring
We currently have many client requests for interesting projects and sometimes have to postpone or decline, because they are too many to handle with our current team size. We also want to work more on developing our own services and technology solutions.
It is an exciting time for us as Hacarus continues to grow. You could be part of our team and experience all the elements of our product development.
Open Position
Data Scientist
Data Scientist Job in KYOTO
HACARUS Inc. is a startup company in Kyoto, providing data science services using AI technology.
Our goal is to "build intelligence to coexist with people".
HACARUS offers a specialized artificial intelligence package called "HACARUS-X", that adopts the method of sparse modeling. Sparse modeling differentiates from current mainstream deep learning in its ability to generate results from small amounts of data. With this technology, we are able to offer "lightweight", "safe", and "explainable" solutions. We currently build solutions with a focus on medical and industrial use cases.
--- AI Solution for Clinical Research · Diagnosis · Medical Treatment Support
HACARUS has been involved in the healthcare field and build expertise since the company was founded in 2014. Using HACARUS analysis technology, we can use basic health parameters such as blood pressure, heart rate, activity level, brain CT, and MRI to support patient diagnosis and treatment. Currently, we participate in a joint medical study with a major international pharmaceutical company.
--- Embedded · FPGA Low Power AI Solution
HACARUS' AI solution is characterized by its ability to operate with low computing resources and low power consumption. Based on this feature, we are developing FPGA chip based AI technology, for easy incorporation into devices such as cameras, routers, and industrial equipment. We are introducing the technology to embedded devices, and we are currently in joint development with router manufacturing companies.
If you have a interest in Hacarus, please contact us on Wantedly!
Thank you,