and accuracy, as well as minimizing our energy, we can utilize it in many aspects of our lives. To achieve this, LeapMind is engineering both the hardware and software from the ground up to provide these platforms and services. We are one of the few companies in the world concurrently carrying out R&D (research & development) and implementation. Moreover, we have been researching deep learning even before it has gained the polarity today for about 7 years now. Currently, we are recognized by large enterprises and corporations from all over the world. Our goal is to create a world where anyone can utilize machine learning like never before, and to make this device accessible from anywhere.
As a Series C About 35 billion JPY of funds raised 2019. 1 9 10 10 2018. 6 Selected for “J-Startup”Ministry of Economy, Trade and Industry’s start-up support measure LeapMind’s unique technology open-source of “Blueoil” 9 In the mechanism by FPGA for Deep Learning, achieved 10.5fps inferred speed in object detection task
Technology Ventures, Intel Capital, Mitsui and Co., Ltd, Toyota Motor Corporation, SBI Investment Co., Ltd., etc. Intel Corporation, Toyota Motor Corporation, Aioi Nissay Dowa Insurance Co., Ltd, NTT Data Co., Ltd, Hakuhodo Co., Ltd, etc. Mitsui Sumitomo Bank, Mizuho Bank, Mitsubishi UFJ Bank, etc.
connection is not required Embedding FPGA・ASIC Cheap device unit price Fast response Low power consumption High device unit price High power consumption Requires the internet Slow response High usage fee To make thing intelligence
affluent society by developing Blueoil, an open source software that creates low-bit compact neural networks, and by designing a unique processor for deep learning that operates with low power consumption. 17
Seamless conversion / optimization of models from minimal quantization to hardware language Server FPGA / CPU C++ API Python API API DNN Inference Engine DNN Accelerator (on FPGA) Trained model configuration .yml Tf-graph .pb Dataset Quantized DNN Inference Accelerator C++ compiler Off-the-shelf Network Models ・Classification ・Object Detection ・Semantic Segmentation .py Model configuration .yml .py Forward Backward Update Quantization-Aware DNN Training Graph format conversion Graph Conversion Generic graph optimization Device-specific optimization / code generation Dataset Preparation Training and Quantization Conversion and Optimization Inference Accelerator binary Closed source FPGA SDK Quartus® Prime Vivado® Accelerator RTL source code .v
ultra-small quantization deep learning With LeapMind's technology, digital cameras and TVs that stopped improving in terms of power, performance and cost, can be installed with AI in mass-produced items such as drive recorders and drones.
and development projects with many companies, including companies that want to introduce driving assistance, device incorporation, factory inspection and abnormality detection, etc., in order to integrate deep learning into our lives. Driving support technology Non-defective product / defective product classification Automatic control support for drones Automatic detection of scratches and cracks Danger detection by surveillance camera Foreign object detection
have limited power and space and don’t have guaranteed continuous connection to the Internet by providing object and scene recognition, performing weather analysis and supporting the control of wipers, lights, etc., I want to use it with a limited power source/space I don’t want to process my data to the cloud I want to process the inference in real time I don’t have internet/ My internet connection is unstable Automotive
and limited space of electrical appliances such as light bulbs and microwave ovens,and can automatically dims the lights and heat dishes up to aid in cooking. Electronics I want to use it with a limited power source/space I don’t want to process my data to the cloud I want to process the inference in real time I don’t have internet/ My internet connection is unstable
improved by constructing a system that detects abnormalities from image data and automating anomaly detection, which was conventionally done by human eyes. Food Inspections I want to use it with a limited power source/space I don’t want to process my data to the cloud I want to process the inference in real time I don’t have internet/ My internet connection is unstable
available everywhere In order to do so, we are developing solutions so that anyone can immediately integrate edge deep learning However, it is often difficult for customers alone to implement. In order to solve this issue, we are developing our project as well as our consulting business
a service that visualized and matched an engineers' skills,and later expanded and transferred his business to Singapore. He founded LeapMind to create a platform easily accessible to anyone as well as to reduce the complex advanced technologies for society, and to advance the world by making deep learning technology “compact and simple”,. CTO Hiroyuki Tokunaga He completed his master's program at the University of Tokyo Graduate School of Information Science and Engineering in 2007. Prior to becoming the Director and CTO of LeapMind in 2018, he worked for Yahoo Japan Co., Ltd., Preferred Infrastructure Co., Ltd. and Smart News Co., Ltd. Chief Research Officer & Chief Scientist Atsunori Kanemura, Ph.D. He holds a Ph.D. in Informatics at Kyoto University, Japan. He has enrolled in research institutions both in Japan and overseas, and has over 50 academic papers, over 100 academic presentations, and has lectured at AIAA's flagship international conference AAAI. In order to present the future image of machines with cognitive functions used in various places through research, he joined LeapMind as Executive Officer CRO in 2018. CFO Shohei Sasaki Prior to becoming the director and CFO of LeapMind in 2018, he was responsible for accounting and finance at Aeria Inc. He then transferred to Acquire Co., Ltd. as CFO and was in charge of buyouts. He then co-founded Cloudworks Inc. as the CFO. CHRO Takeo Sawada He obtained his master's degree in computer science at the University of Tokyo Graduate School of Information Science and Technology in 2013. Prior to his role as LeapMind Executive Officer, VP of Engineering in 2018, he worked as SRE at Google (Mountain View, CA) in 2013 and Google (New York City, NY) in 2014. In 2017, he worked as an SRE at Dropbox (San Francisco, CA).
・Members come from 14 countries ・English and Japanese is used interchangeably, and meetings are often in English ・Members have different backgrounds:Entrepreneur, researcher, consultant, etc.、 Diversity ・Flex system and flexible work style ・Flat organization. Close relationship between board member and employees. Freedom Characteristics of LeapMind
about deep learning and other various topics with in-house experts Every 4 months, all employees participate in an activity of their choosing that challenges them for a period of 5 days External study sessions Held irregularly on various topics LeapMind BLOG Stories and posts are made by employees Company events HackDays
to create a system to support employee’s diﬀerent working styles From time to time, we consider introducing a new system depending on the phases! Work Style Flex system, remote work, and flexible work styles such as “refresh” leave Free Drinks Mineral water and coffee on the house Education Support Support for expenses required to improve skills, such as for taking Coursera courses and purchasing books Club Activity Activity expense support to increase and encourage exchange opportunities through common interests 1 on 1 We conduct individual interviews and encourage open communication to support employee growth Office Event Meetups and barbecues to contribute information and foster communication among all employees Development Environment Support to improve the development environment. Such as the supply of 4K monitors, laptops and high-back chairs Fun Budget Expense support for team building. Such as for hosting lunch parties and events
to improve the employee experience to support those who are coming from abroad to join LeapMind Support in obtaining a work visa in Japan Partial expense covered for those moving from abroad Online interviews for those who live far away English support channel for personal inquiries Internal materials are written in both English and Japanese Full expenses covered for Japanese Language Proficiency Test JLPT (N 1, 2) Aあ
Regardless of age, gender, nationality, race, etc. We, at LeapMind, want to work with someone who wants to make the most out of their skills and abilities, and also with someone who can respect diversity. Our backgrounds aside, we all share a common goal. To create innovative devices with machine learning and make them available everywhere We are doing the most exciting things right now at Leapmind. Would you like to join us in achieving LeapMind’s vision?