Posts on Mar 2019

Call for Applications for PFN Summer Internship 2019 in Tokyo

The application for PFN Summer Internship 2019 was closed.
Thank you for many applications.

Preferred Networks (PFN) is looking for enthusiastic interns who can work with us in our Tokyo office this summer. Students who applied for previous internships are also eligible to re-apply (except those who previously interned with PFN). We welcome students who want to develop new technologies, software, and services across a wide range of computer science areas including machine learning with us.


Important notice:

Note: This program is only for students who already have visa eligibility to work in Japan as an intern this summer. We are not accepting applications from students who need support for obtaining the designated activity visa to work as an intern because the due date for processing such applications has already passed.



Characteristics of PFN Internship

  • Over the two-month period, PFN researchers and engineers will be assigned to work with you as a mentor. You will have opportunities to discuss, study, and develop in your theme with specialists in various fields including deep learning, computer vision, natural language processing, robotics, bio-healthcare, reinforcement learning, and distributed processing.
  • After the internship, you can make your research result public by writing a paper or making it OSS, to the extent to which is possible.  



  • Start date: Early August, depends on your schedule
  • End date: Sept. 20 (Fri), 2019
    • You can choose to continue to work in the week of Sept. 24-27 under the same terms and conditions.
      • This year’s internship will end on Sept. 20 in principle in consideration of the fact that many schools start their fall semester in late September.
      • If your school is not the case stated above and you need more time to finalize your research or want to spend more time with our staff, you can extend your internship until Sept. 27 under the same terms and conditions.
      • We understand you may have to be absent for e.g. lab activities, attending academic conferences, returning home for family commitments. We are flexible about your need to take day-off due to these reasons.

Theme List

  • For 2019 Domestic Internship at PFN, we prepared the themes listed further below.
  • During the selection process or after completion, you will discuss with our members to finalize your theme. You must at least choose your 1st and 2nd preferences in the application form. If you have more than two areas of interest, select a 3rd preference, which is optional.



  1. Theoretical research of ML/DL
  2. Computer vision for 3D scene
    • Neural networks for 3D tasks (differentiable renderer, 3D reconstruction using neural network)
    • Development of SLAM and 3D reconstruction
    • Visual-SLAM
  3. Computer vision on time-series dataset
    • Video analytics (sports, etc…)
  4. Computer vision for other general topics
    • Object detection
    • Segmentation
    • Image classification
    • Few-shot learning
    • Image generation
  5. Application of deep learning to Anime / Supportive software for creators
  6. Reinforcement learning
  7. Research and development on applications of machine learning algorithms such as mathematical optimization, simulation, and time-series prediction
  8. Bio-healthcare
  9. Chemoinformatics / Materials Informatics
  10. Dialog, semantic parsing, symbol grounding, reasoning, or translation
  11. Speech and signal processing
  12. Interface and interaction
    • VR or AR
    • Human Computer Interaction / Human Machine Interaction, Human Robot Interaction
  13. Robotics – DL/DRL for robotics
  14. Robotics – Research and development of robot simulation
  15. Robotics – Planning of mobile robots
  16. Development of Chainer or libraries on Chainer
    • Development of Chainer
    • Development of CuPy
    • Development of area-specific libraries based on Chainer (such as ChainerCV, ChainerRL, ChainerChemistry and ChainerUI)
  17. Research and development on performance optimization of machine learning, etc.
    • Optimization of NN models for inference
    • Development of compiler techniques for deep learning
    • Development of application software for our in-house developed accelerators
    • Development of processor, low-power computer architecture and VLSI technology for deep learning
  18. Research and development of infrastructure for machine learning
    • HPC and distributed data management for distributed deep learning / deep learning
    • Development of experiment management systems, cluster management, experimental environment optimization for machine learning
    • Research and development of Edge Heavy Computing/In-Network Computing
    • Development of Optuna
  19. Research and development of pipeline automation tools for machine learning
    • Topics related to automation of machine learning pipeline, including hyper parameter optimization, architecture search, and feature engineering
  20. Front-end development
    • 【Renewal on April 5th】Development of information visualization tool, annotation tool / front-end for machine learning
    • Development of information visualization tool / front-end for machine learning
  21. Product Design
    • Concept design, sketch and user interaction for robot products
    • Development of 3D-CAD models and hardware prototypes
  22. Others
    • Free topic. We also welcome self motivated applicants who can take the lead of their own research project.



Internship Location

PFN Tokyo Office

Otemachi Bldg., 1-6-1, Otemachi, Chiyoda-ku, Tokyo, Japan 100-0004



Key Qualifications

  • PFN is seeking highly motivated and skillful individuals who can develop applications, tools, etc. independently.
  • If you have knowledge or development experience in the themes that are stated in【Theme List】section, we’ll take them into consideration, but it’s not required. The minimum requirements are:
    • Currently enrolled in high school, technical college, university, or graduate school. Negotiable for those attending other higher education institutions
    • Fluent in Japanese or English
    • Strong communication skills
    • Prior experience in programming (any language)
    • Can work from our Tokyo office on weekdays
      (we’re not accepting remote working for the internship program currently)
  • Do not hesitate to apply even if you don’t have prior experience in full-scale development.



Important Notes before You Apply

  • We are not accepting applications from students who need support for obtaining the designated activity visa to work as an intern. Call for such applications for this year has already been closed.
  • Non-Japanese students who are studying at Japanese universities with student visa must make sure to get 【Permission for Other Activity / 資格外活動許可】before the start date of your PFN internship program
  • You need to let us know in advance if any administrative work is required to receive academic credit from your school. Please note that depending on the complexity of the work, PFN may not be able to accommodate your request.



How to Apply

  • Click this application form to apply. To access the application form, you will need to log in with a Google account.  We use your personal information filled in the application form for selection of summer interns.
  • About the submission of your portfolio
    • Summarize your skills and qualifications freely in an A4 paper to pitch yourself.
    • Highlight and showcase some of your best work such as software you have developed, published papers, awards or prizes you have received, programming contests you have participated in, your blog, your sites, Twitter account, and other social media sites.
    • Please upload it using the “Portfolio to pitch yourself” field in the application form.
  • Deadline:April 18th (Thu) 12:00 PM JST (No late submission allowed)
  • For inquiries:email to



Selection Process

  • First screening
    • After we close the application on April 18, we will send the coding tests to all applicants (in principle) on April 19.
      【Updated on Mar 29】For those who choose 21. Product Design, we will send you the guideline of a task instead of coding test.
    • The deadline for completing these tests is May 7 (subject to change).
  • Interview
    • The interview will take place within 3 weeks from May 27.
    • For students living in distant areas, PFN will arrange an online (video) interview via Wepow.
  • Offer
    • The letter of acceptance is planned to be delivered on June 25 or later



Basic Employment Conditions And Benefits

  • Salaries:
    • 2,500 yen an hour for technical college, university, graduate school students
    • 2,000 yen an hour for high school students
  • Work hours:8 hours in principle. 5 days a week excluding Saturdays, Sundays, and public holidays.  
  • Commuting fee support:PFN will pay for your daily commute to and from the office in an approved route.
  • Travel cost:For students traveling a long distance by plane or Shinkansen to participate in the internship, PFN will support the cost for one round trip to relocate between the Tokyo area and the place where you’re currently living.
  • Accommodation support:For students coming from distant areas, PFN will provide an accommodation allowance of 5,000 yen per day for the entire period of your internship including holidays.
    • You need to arrange a place to stay by yourself. Reasonable weekly rental apartments are available near PFN office ranging from 100,000 to 150,000 yen a month.
    • Please note that the accommodation allowance is taxable.

Preferred Networks builds MN-2, a state-of-the-art supercomputer powered with NVIDIA GPUs.

It will become operational in July to provide a combined computing power of 200*1 PetaFLOPS*2.


March 18, 2019, Tokyo Japan – Preferred Networks, Inc. (PFN, Head Office: Tokyo, President & CEO: Toru Nishikawa) will independently build a new private supercomputer called MN-2 and start operating it in July 2019.

MN-2 is a cutting-edge multi-node GPGPU*3 computing platform, using NVIDIA(R) V100 Tensor Core GPUs. This, combined with two other PFN private supercomputers ― MN-1 (in operation since September 2017) and MN-1b (in operation since July 2018), will provide PFN with total computing resources of about 200 PetaFLOPS. PFN also plans to start operating MN-3, a private supercomputer with PFN’s proprietary deep learning processor MN-Core(TM), in spring 2020.

By continuing to invest in computing resources, PFN will further accelerate practical applications of research and development in deep learning technologies and establish a competitive edge in the global development race.

Conceptual image of the completed MN-2


  • Outline of PFN’s next-generation private supercomputer MN-2

MN-2 is PFN’s private supercomputer equipped with 5,760 latest CPU cores as well as 1,024 NVIDIA V100 Tensor Core GPUs and will be fully operational in July 2019. MN-2 is to be built on the premises of Yokohama Institute for Earth Sciences, Japan Agency for Marine-Earth Science and Technology. MN-2 will not only work with MN-3, which is scheduled to start operation in 2020 on the same site, but also connect with MN-1 and MN-1b, MN-2’s predecessors that are currently up and running, in a closed network. MN-2 can theoretically perform about 128 PetaFLOPS in mixed precision calculations, a method used in deep learning. This means that MN-2 alone has more than double the peak performance of MN-1b.

Each node on MN-2 has four 100-gigabit Ethernets, in conjunction with the adoption of RoCEv2*4, to interconnect with other GPU nodes. The uniquely tuned interconnect realizes high-speed, multi-node processing. Concurrently, PFN will self-build software-defined storage*5 with a total capacity of over 10PB and optimize data access in machine learning to speed up the training process.

PFN will fully utilize the open-source deep learning framework Chainer(TM) on MN-2 to further accelerate research and development in fields that require a large amount of computing resources such as personal robots, transportation systems, manufacturing, bio/healthcare, sports, and creative industries.

Comments from Takuya Akiba,
Corporate Officer, VP of Systems, Preferred Networks, Inc.

“We have been utilizing large-scale data centers with the state-of-the-art NVIDIA GPUs to do research and development on deep learning technology and its applications. High computational power is one of the major pillars of deep learning R&D. We are confident that the MN-2 with 1,024 NVIDIA V100s will further accelerate our R&D.”


Comments from Masataka Osaki,
Japan Country Manager, Vice President of Corporate Sales, NVIDIA

“NVIDIA is truly honored that Preferred Networks has chosen NVIDIA V100 for the MN-2, in addition to the currently operating MN-1 and MN-1b, also powered with our cutting-edge GPUs for data centers. We anticipate that the MN-2, accelerated by NVIDIA’s flagship product with high-speed GPU interconnect NVLink, will spur R&D of deep learning technologies and produce world-leading solutions.”


*1: The figure for MN-1 is the total PetaFLOPS in half precision. For MN-1b and MN-2, the figures are PetaFLOPS in mixed precisions. Mixed precisions are the combined use of more than one precision formats of floating-point operations.

*2: PetaFLOPS is a unit measuring computer performance. Peta is 1,000 trillion (10 to the power of 15) and FLOPS is used to count floating-point operations per second. Therefore, 1 PetaFLOPS means that a computer is capable of performing 1,000 trillion floating-point calculations per second.

*3: General-purpose computing on GPU

*4: RDMA over Converged Ethernet. RoCEv2 is one of the network protocols for direct memory access between remote nodes (RDMA) and a method to achieve low latency and high throughput on the Ethernet.

*5: Software-defined storage is a storage system in which software is used to centrally control distributed data storages and increase their utilization ratios.


*MN-Core(TM) and Chainer(TM) are the trademarks or registered trademarks of Preferred Networks, Inc. in Japan and elsewhere.