Researcher / Research Engineer (Remote Sensor Data and Machine Learning)
Preferred Networks, Inc., Japan
Job Description
Preferred Networks NICT()(https://www.nict.go.jp/press/2017/11/29-1.html)(https://www.nict.go.jp/press/2022/01/25-1.html)(NICT
- DNN compression
- 1-2
- (SAR) 1
PE/
At Preferred Networks, we are conducting a series of research projects employing the massive remote sensor data from advanced weather radars (https://www.nict.go.jp/press/2017/11/29-1.html) and aerial radiometric landscape imaging sensor (https://www.nict.go.jp/press/2022/01/25-1.html). The projects are co-operated with NICT, funded by the Ministry of Inner Affairs and Communications of Japan.
We are looking for graduated students or undergraduate students who can conduct one of several projects passionately as part-time engineers and researchers.
Research projects include:
- Research and development on the DNN lossy compression of the advanced high-spatio-temporal resolution 3D weather radar data 1 person
- Research and development on the estimation or prediction of various weather conditions using the advanced high-spatio-temporal resolution 3D weather radar data 1 or 2 persons
- Research and development on the detection and classification of several land surface classes based on the massive collection of SAR images and corresponding GIS dataset 1 Person
Regardless of the subject chosen, a candidate(s) will conduct the following tasks by necessity basis, jointly working with FTEs, other PEs and internship students:
- Preparation and maintenance of the dataset
- Maintenance and improvement of the computation environment
- Development and maintenance of our in-house program library for this project
- Attending discussions and meetings with stakeholders in-and-out of the company
- Conduct massive computational experiments using the in-house GPU computing cluster
- Writing and publishing academic papers and patents, presenting in appropriate locations
Location
- Remote
- PFN Office (Otemachi, Tokyo) when necessary
Employment period
- 6~
- 6 months~
Requirements
Key qualifications:
- Python 3, numpy/scipy, pytorch
- Git/Github
- Linux
- Self-motivated in pursuing goals.
- Business-level communication skills, either in English or Japanese.
- 1+ year of experience in development with both Python 3, numpy/scipy, and pytorch.
- Familiarity with Git and GitHub-based development flow.
- Familiarity with Ramp;D activities in Linux environments.
The following experiences will set you apart from other candidates:
- GIS(GDAL, Geopandas, Geojson, QGIS)
- ,
- GPU
- Docker/kubernetes
- Experiences dealing with GIS data and related libraries, softwares such as GDAL, Geopandas, QGIS etc.
- Strongly knowledgeable about, and/or having strong Ramp;D experiences on computer vision, in particular, semantic segmentation
- Strongly knowledgeable about, and/or having strong Ramp;D experiences on data assimilation and time-series analysis such as ARMA and Kalman filters
- Strongly knowledgeable about, and/or having strong Ramp;D experiences on pretrained large machine learning models, multi-task learning, and meta-learning.
- Strongly knowledgeable about, and/or having strong Ramp;D experiences on whether radar data processing
- Strongly knowledgeable about, and/or having strong Ramp;D experiences on machine learning and AI-related techniques for SAR images and optical images of satellites/airbornes
- Knowledgeable about physics of whether and atmosphere
- Mathematical skill to formulate and approximate physical phenomenon into
- Eager to start Ramp;D activities from collecting and organizing your own dataset
- Development experience of multi-GPU environment
- Development experience with Docker/kubernetes environment
- Bachelor’s Degree, Master’s Degree, or Ph.D. in Computer Science or related field.
- Prove publication record to academic journals and peer-reviews International Conference as the 1st author.
Contents in a resume
- /
- GitHub
- The format of your resume does not need to follow any particular one. To help us understand as much as possible about you, please write about projects you have participated in, part-time jobs/internships, papers you have written, implementations you have published, etc.
- Please include links to your own technical blogs, GitHub repositories, and papers, if any.
Benefits
- : 10
- : 2500
- Working hours: 10 hours or more per week (However, working days or times can be decided flexibly)
- Hourly wage: 2500 yen