Embedded ML Platform Vendor Edge Impulse Secures $34M in New Funding
Edge Impulse, which offers a SaaS-based machine learning platform for creating projects for edge devices, has received $34 million in new Series B funding to grow its operations, marketing and staff.
The two-year-old startup announced the additional funding, which was led by Coatue, on Dec. 9 (Thursday). Also joining the latest funding round are existing investors Canaan Partners, Acrew Capital, Fika Ventures, Momenta Ventures and Knollwood Investment Advisory.
The Edge Impulse SaaS platform is built to provide a low-code platform for enterprise software developers, engineers and domain experts to use ML for embedded applications, without requiring deep knowledge of ML. The platform is also designed to speed up development time from years to weeks, according to the company. Edge Impulse is aimed at industrial and professional ML projects including predictive maintenance, asset tracking and monitoring and human and animal sensing.
“Edge Impulse gives embedded engineers the possibility to ingest sensor data, and to draw conclusions from that,” Jan Jongboom, the co-founder and chief technology officer of the company, told EnterpriseAI. That includes sensors on machines that monitor for sounds, vibrations, abnormal system temperatures and other factors, he said.
“It could be a microphone that is listening, for example, for noises in industrial steam pipes,” he said. “Sometimes we see pipes where one of the valves is not properly open, so you can listen for that.”
Other uses also include wearables for health monitoring, sensors for monitoring animal poachers in Africa and activity detection applications for mobile devices, said Jongboom.
The Edge Impulse SaaS platform include all the steps needed to build these types of models, such as data collection, data processing, signal processing and deploying the data to the sensor, he said.
The platform is available for free use to individual developers, who have created more than 50,000 ML projects using the platform so far, or through a paid service for enterprise customers, he said.
For customers, Edge Impulse can help them see what data they have and find ways to use it that never imagined in the past, said Jongboom. Many companies are sitting on sensor data, but if they do not know what it is, then it is hard to draw conclusions from it, he added.
“With our enterprise customers, we have a solution engineering team that goes in with the customer, learns as much as they can about the type of data they have and the problems that they are facing, then we will help them make that data insightful and integrated,” said Jongboom.
Customers can go through the entire process using a low-code approach or they can hand it over to their data scientists who can dive deeper into the platform’s code and perform more tasks, he said.
Many enterprises have been experimenting with such projects, he said, but have been looking for commercially available products like this to make it easier for them to do on their own.
“We launched early on, and we are developing the market in unison with the growth of the company,” said Jongboom.
The $34 million in new funding will be used to double Edge Impulse’s staff from 40 people to 80 employees through 2022, particularly in software engineering, he said. “First and foremost, we want to continue the kind of the trajectory where we grow both the community and our enterprise customers at the same time, and this will just accelerate that.”
One of the most interesting things the company is working on today are modeling tools that will better help users make decisions from their data when running models, he said.
The tools take raw data from the sensor and can show when a monitored event is detected so it can be reviewed by a user who can then learn how the model will perform in real life, said Jongboom. “I think this is going to be really interesting and I think this is a trend that we are going to see outside of just sensor data, but also with normal ML models.”
Since the launch of Edge Impulse in mid-2019, the company said it has almost 30,000 developers from thousands of enterprises who are using the platform. Some enterprises using the platform include Oura, Polycom, Advantech and NASA. The company said it aims to help customers use machine learning on embedded devices for sensors, audio and computer vision at scale.
The no-code trend in software platforms is becoming more popular. AWS recently announced two new members of its SageMaker machine learning training platform family – Amazon SageMaker Canvas, which will allow non-developers to create no-code ML projects – and a free, public preview of its all-new Amazon SageMaker Studio Lab, which will allow anyone interested in learning more about ML to experiment with the technology without even needing an AWS account.
In May, Google Cloud unveiled its Vertex AI managed ML platform, which is designed to streamline and accelerate ML modeling and maintenance to help overwhelmed enterprises get a better handle of their ML and AI initiatives.