Akridata announced the launch of the Akridata Edge Data Platform™, which creates and manages smart data pipelines and AI workflows spanning Edge-Core-Cloud resources. The Akridata software solves the problems that emerge when streams of rich data from physically scattered Edge devices create an avalanche of data that is impossible to organize, filter, access, and process. It is now common for organizations to collect tens of Terabytes of data per day from a single autonomous device.
With the AI infrastructure platform in the Data-Centric AI category, Akridata provides a decentralized structure and scalable process to deliver curated, consistent, and relevant AI data sets. Akridata was formed in 2018 to solve AI’s data problem. Solving this Exascale-class problem is a requirement to move AI out of experimentation and into real-world production. From automotive and transportation to retail and healthcare, this need is a major obstacle that prevents AI-enabled products from making it to the marketplace.
“The diverse requirements of ADAS/AV (Advanced Driver Assisted Systems/Autonomous Vehicle) necessitate many elements including deep learning, cloud deployment and in-vehicle services, among others. What ties all these together is data and a massive data challenge,” said Kishore Jonnalagedda, Director of Engineering, Toyota Motor Company North America. “Akridata brings us a comprehensive and novel solution that drives efficiencies, lowers the cost, and accelerates our team’s work toward our objectives. We will gain immediate leverage by automating data pipelines from edge locations to the cloud, allowing our data science and product development teams to focus on what matters most: delivering best-in-class ADAS/AV solutions and providing mobility for all.”
The Akridata solution is optimized for advanced AI workloads, providing built-in capabilities for AI-oriented data organization, transformation, and filtering tasks. It allows tracing and tracking of data from inception to inference, it enables traceable AI and broadly complements industry efforts towards explainable AI (XAI). This makes it possible to track the evolution of AI models and link the behavior of AI models in the field to the data sets that contributed to the specific model used by a specific device or service.
“Akridata is making the autonomous world possible by providing the last piece of the puzzle: an integrated Edge-Core-Cloud Data Platform that solves the data problem at the heart of all real-world AI systems,” said Kumar Ganapathy, co-founder, and CEO of Akridata. “The future of AI is all about data, and our focus on AI data since inception gives Akridata a first-mover advantage. We are pleased to launch the first infrastructure product in the Data-Centric AI Category, and to be working with a range of customers including industry leaders like Toyota Motor Company North America.”
“To thrive in the emerging global IT infrastructure, companies and other organizations will need to exploit heterogeneous data from highly distributed sources ranging from edge devices to powerful computers in clouds and data centers,” said Steve Conway, Senior Advisor at Hyperion Research. “Akridata is well-positioned to benefit from the strong growth Hyperion Research expects in this emerging data-centric market.”
Akridata’s innovative new solution enables the integration of Deep Learning with Inference, Edge Commerce, Data Governance, and enterprise applications. It was specifically developed to address the Exascale-class data challenges that come with delivering advanced AI, autonomous devices, and unattended services.
“Advanced AI models are increasingly created and run on the cloud, but need good quality data from edge devices,” said Jon Jones, Director – Go-to-Market for AI/ML, EC2, and Autonomous Vehicles, AWS. “The high volume and complex nature of this data has created a new exascale-class problem. Akridata is addressing the problem by managing smart pipelines for ingesting, filtering, curating, tracking, and staging of AI data.”
AI Data Complexities
The autonomous world depends on continuous Deep Learning using large volumes of complex AI data sets. Streams of rich data – such as video and lidar data – generated by fixed or mobile edge devices must be organized, filtered, tracked and processed across the Edge-Core-Cloud resources. Massive amounts of data are being generated at the Edge by these devices. For example, a self-driving car in its test phase can generate multiple Terabytes of data in a single day. By 2025, 75 percent of the 175 Zettabytes of new data generated will come from the Edge, according to industry experts.
“Especially for newer workloads like AD/ADAS development, HPC/AI workflows are increasingly grappling with an Exascale-class data challenge. The continuous flow of data from intelligent edge devices into the cloud creates significant demand for data curation as it’s absolutely required for subsequent AI & software development and validation pipelines,” said Kurt Niebuhr, Principal Program Manager, HPC/AI Ecosystem and Workload Incubation, at Microsoft Azure. “Data-Centric AI solutions such as Akridata’s Edge Data Platform fill that need and help customers to match the sophistication of their analytics and models with high-quality and relevant data.”
The smart Akridata Edge Data Platform is distributed and helps optimize the processing, storage, and movement of data across the Edge, the Core, and the Cloud. The Akridata platform is available and has been shown to deliver ten times faster time-to-access the right data, four times more efficient usage of compute and storage, and two times better productivity for data scientists and Machine Learning engineers. With Akridata’s solution, real-world AI-enabled products can be brought to life.
