WHAT WE DO
Inspirit IoT's vision is to be the global leader in delivering sensor intelligence.
We take pride in our mission to provide hardware accelerator solutions coupled with an advanced high-level synthesis and machine-learning optimization toolkit to enable smart IoT applications.
Inspirit IoT delivers a compelling channel and customer experience in markets where performance, power, and time-to-market are a mission or regulatory imperative. As one team, we work hard to be intimately aware of, and responsive to, our customers needs.
The experts in Sensor Intelligence
WHO WE ARE
We are a pragmatic, diverse, and international team, empowering only the best talent in an un-bureaucratic environment, where our team are valued and rewarded for “challenging the status quo”.
We thrive on making an impact. By making the difference to the world’s most challenging hardware problems, we can be proud of the societal value that we create.
PARTNERS & SUPPORT
STRATEGIC ALLIANCE PARTNER
Inspirit IoT's advanced design tools and design services help Amazon customers meet their specific computing need in either AWS FPGA cloud or IoT edge devices.
Intel IoT Solutions Allliance Partner
The Intel IoT Solution Alliance accelerates the design and deployment of intelligent devices and analytics. Inspirit IoT is partnering with Intel and its OEM Partners to develop and deliver first-in-market IoT solutions.
STRATEGIC ALLIANCE PARTNER
Inspirit IoT embeds low-power Intel FPGA chips, which can be reconfigured to target various smart sound IoT applications.
STRATEGIC ALLIANCE PARTNER
Inspirit IoT provides DNN Optimization and HLS design solutions for IBM customers targeting machine learning applications.
Siemens Frontier Program Partner
STRATEGIC ALLIANCE PARTNER
Inspirit IoT is working with Siemens to help transform robotics: reduce time to market, enhance flexibility, and increase efficiency for their customers.
Advanced Digital Sciences Center
Inspirit IoT was forged as an R&D project to solve the semiconductor industry challenge: design productivity greatly trails the silicon capacity driven by the Moore's Law.
National Science Foundation
Inspirit IoT is awarded NSF SBIR seed funding based on our technology’s innovativeness, commercial potential, and possible societal impact.
Inspirit IoT is awarded seed funding and is incubated in I-Start Entrepreneur Assistance Program at Research Park of University of Illinois at Urbana-Champaign.
ACCELERATING MACHINE LEARNING
Performance Made Easy
Machine learning delivers application opportunities and competitive advantage yet is compute intensive, requiring acceleration to meet real-time performance and power-efficiency constraints. However, design and deployment of hardware accelerators require detailed design expertise.
We Speak Your Language
Machine learning solutions are designed in TensorFlow, Caffe, CNTK, Python, R, and Matlab. Inspirit IoT lets you design your machine learning application using languages and tools data scientists are familiar with, automating the translation to detailed hardware accelerators that give the performance you need.
Your Network, Your Solution
Machine learning is rapidly evolving -- whether you want to deploy the latest popular network architecture, design your own from scratch or anything in between, we'll help get you there. Inspirit IoT's tools and libraries explore implementation options to give you the best speed performance for your network.
Performance Where You Need It
Not all applications are built the same, and neither should the accelerators. Some use cases are well served for cloud deployment to reduce infrastructure costs, while security, proprietary information, privacy regulations, or existing infrastructure requirements may require in-house clouds in others. With expensive communications or location constraints, your machine learning may need to be in an edge device; Inspirit IoT helps deliver machine learning performance where you need it, whether the cloud, private servers or in edge devices.
Power, Performance, Time to Market
Whether your challenge is meeting latency or throughput constraints, minimizing power and energy consumption to reduce electricity and cooling cost of data centers or lengthen the battery life for edge devices, or delivering a product to market before your competitors, Inspirit IoT is there to deliver. Inspirit IoT's FPGA as a Service capability, tools and libraries can explore performance and power tradeoffs automatically, helping you to meet your application goals while delivering a solution to the market faster than prior design methods.
Inspirit IoT's HLS tools include a state-of-the-art general-purpose HLS core with extensive IP integration in order to integrate IPs for pipelined functional units, floating-point units, third-party IPs, and customer IPs. Paired with our library of ML components Inspirit IoT's HLS tools provide ML domain-specific synthesis.
Inspirit IoT HLS currently outperforms market competitors 2 x based on Quality of Results.
AUDIO EVENT DETECTION
Inspirit IoT's library of Smart Sound audio-specific IPs includes localization, audio enhancement, event detection (e.g. screaming, gunshots), and keyword detection.
Available on Lattice and Intel FPGAs.
Inspirit's smart soundboard can be deployed indoors or outdoors across a wide range of events, in any non-military application.
Inspirit IoT's TensorFlow (TM) - based framework can train networks, explore reductions in weight bitwidth, clustering and pruning of weights in order to reduce size and complexity of implementations, and retraining networks to retain the expected quality of results.
The only tool in the AI market that delivers 10x better performance compared to the initial Deep Neural Network targeting different IoT devices, such as smart phones, SoC chips, and FPGAs.
Inspirit IoT was incubated in 2011 on the back of a research grant to the Advanced Digital Sciences Center of Singapore – a collaboration of the Singapore Government’s A*Star Research and the University of Illinois at Urbana-Champaign (UIUC). The research aimed to address the needs of creating energy-efficient hardware accelerators of computation-intensive algorithms in an automated way as well as higher hardware design productivity and quality.
Our award-winning technology was commercialized in June 2016 at EnterpriseWorks at the Research Park of UIUC. Meanwhile, Inspirit IoT, Inc was awarded a National Science Foundation (NSF) Small Business Innovation Research (SBIR) grant and UIUC I-Start Accelerator grant to conduct research and development (R&D) on automating design and deployment of hardware for accelerated machine learning.
We are now delivering POCs with the world's largest companies, OEM partners, and System Integrators. Inspirit IoT is a registered Intel Technology Provider, AWS Technology Partner, IBM Business Partner, and Siemens Frontier Program Partner.
The company currently employs 10 employees and has operations in United States, Singapore, and India.
President, Chairman of Board of Directors
Professor, UIUC ECE, PhD UCLA
Chief Technology Officer
PhD ECE, UW-Madison
Chief Innovation Officer
Executive MBA TRIUM (NYU, HEC, LSE)
PhD CE, Univ. Alabama-Huntsville
Senior Software Engineer
PhD Electrical and Computer Engineering, Illinois
Board of Directors
Professor, UCLA CS, PhD CS, UIUC
Co-founder: Aplus, AutoESL, Neptune, Falcon Computing
Board of Directors
Professor ECE UIUC, PhD CS, UC-Berkeley
Co-founder & CTO MulticoreWare
Co-founder & Board-member Personify
PhD Accounting, CMU
Asst. Prof. Cornell, PhD CS UCLA
AWARD WINNING RESEARCH
Inspirit IoT has contributed to the following research, including two Best Paper Awards, in the domains of IoT, High-Level Synthesis, FPGAs and Artificial Intelligence. Inspirit IoT welcomes industrial research partnerships for the accelerated development of frontier technologies.
Dr. Deming Chen (Chairman and President of Inspirit IoT) and Inspirit IoT intern Xinheng Liu recently won a best paper award at the 2018 IEEE/ACM International Workshop on System-Level Interconnect Prediction (SLIP). This paper outlines a backward pipeline scheduling idea that effectively enables layer overlapping across multiple layers of DNN and reduces latency significantly compared to the conventional pipelining solution.
BEST PAPER AWARD
IN THE NEWS
Sixty-one team participated in the contest to detect minuscule objects in a video shot from a UAV with the highest accuracy and lowest energy consumption. Dr. Deming Chen (Chairman and CEO of Inspirit IoT) and Dr. Wen-mei Hwu (Board Member of Inspirit IoT), along with Illinois postdoc researcher and graduate students Cong Hao, Yuhong Li, Sitao Huang, Xiaofan Zhang, Tianqi Gao, and IBM’s Jinjun Xiong, Inspirit IoT’s Kyle Rupnow, and Boeing’s Huafeng Yu collaborated on the effort.
25 June 2018
San Francisco, California
By accessing the INSPIRIT IoT Inc web pages you agree to the following terms. If you do not agree to the following terms, please notice that you are not allowed to use the site. The contents of INSPIRIT IoT Inc web pages are Copyright © INSPIRIT IoT Inc 2016. Any rights not expressly granted herein are reserved. Reproduction, transfer, distribution or storage of part or all of the contents in any form without the prior written permission of INSPIRIT IoT Inc is prohibited except in accordance with the terms outlined in this document. INSPIRIT IoT Inc provides access to this site free of charge to provide information about INSPIRIT IoT Inc's products and services. The sole authorised use of the site is to obtain information about INSPIRIT IoT Inc's product and service range. No other use is permitted. INSPIRIT IoT Inc consents to you browsing our web pages on your computer or printing copies of extracts from these pages for your personal use only and not for redistribution unless consented to in writing by INSPIRIT IoT Inc. Individual documents in our web pages may be subject to additional terms indicated in those documents. The use of this site and the content therein, is permitted to private, non-commercial use. The use of press releases and other documents classified as public is permitted in public communications if the source for the information has been stated. INSPIRIT IoT Inc assumes no responsibility for the security of this site or your communications with the site. This site is offered AS IS and without warranties of any kind. INSPIRIT IOT Inc disclaims the implied warranties of merchantability and fitness for a specific purpose as well as title or non-infringement. We are not responsible for timeliness, accuracy, unavailability or interruptions in availability, viruses or other defects in the site or its contents. In no event shall INSPIRIT IoT Inc be responsible for any damages to users or their computer systems or otherwise, even if INSPIRIT IoT Inc has been informed of the possibility of such damages and without regard to negligence. For your convenience, INSPIRIT IoT Inc may include links to sites on the web that are owned or operated by third parties. By linking to such third-party site, you agree that INSPIRIT IoT Inc has no control over the content of that site and cannot assume any responsibility for material created or published by such third-party sites. Your access to this site should not be construed as granting, by implication, estoppel or otherwise, any license or right to use any marks or logos appearing on the site without the prior written consent of INSPIRIT IoT Inc or the third party owner thereof. ANTI-BRIBERY POLICY At Inspirit IoT, we strive to achieve superior financial results while adopting high ethical principles. We comply with all applicable laws relating to anti-bribery and corruption in the jurisdictions in which we operate and in particular with respect to the US Foreign Corrupt Practices Act. We expect all employees to conduct themselves with high standards of integrity. We provide training and guidance to employees on our anti-bribery policy and we report gifts and entertainment in accordance with established guidelines. Copyright © INSPIRIT IoT Inc 2016. All rights reserved.