The startup, Run.ai, is the first to bring advanced virtualization to AI chipsets to make training and inference systems faster and cheaper. The Series C round was led by Tiger Global Management and Insight Partners. Other investors include S Capital VC and TLV Partners. Its co-founder and CEO, Lonne Jaffe, was formerly the CEO of Precisely and Syncsort.
Scale accelerates development of computer vision applications
Scale is a cloud-based computer vision API tool from Microsoft that helps developers develop and deploy computer vision applications. It offers access to advanced algorithms for analyzing visual content based on inputs and user choices. It also helps machine learning teams generate high-quality ground-truth data by providing advanced image, NLP, and LiDAR annotation APIs. By automating this step, machine learning teams can focus on building differentiated models worddocx.
Computer vision solutions can also be deployed on local hardware. While this works well for small-scale projects, it can be costly and limited in scalability. Alternatively, edge devices can be combined with peripheral cameras that do not have internet connectivity. By combining the capabilities of these two methods, developers can deploy computer vision applications faster than ever before.
The first step in developing a computer vision application is to build a compelling business case. This includes analyzing the current technical maturity, the availability of labeled data, and the costs of model development, training, and deployment. Additionally, the use cases for the project should be carefully examined for their viability.
In the healthcare sector, computer vision applications can help doctors identify illnesses faster and reduce burnout. For example, medical professionals can use computer vision to analyze patients’ movements and make diagnoses more accurately. Similarly, in the medical field, computer vision applications are used to evaluate the skills of expert learners on self-learning platforms. Furthermore, they can be used to assess surgical students’ skill level. By providing meaningful feedback, computer vision applications can assist individuals in improving their skills.
Run:AI abstracts AI workloads away from GPUs
Run:AI is a company that has developed a software layer that abstracts AI workloads away from GPUs. The idea is that this layer will speed up machine learning workload execution by removing the need to interact with the hardware. The software layer is built to support a wide variety of AI workloads and is compatible with all major cloud providers and on-premise deployments.
This technology provides a flexible, scalable, and secure way to orchestrate AI compute infrastructure. Businesses can use cloud-based or on-premises infrastructure, and can share resources across teams. IT teams can also have better visibility into compute resources. Users from various industries have benefited from Run:AI’s AI platform.
Run:AI supports Nvidia GPUs. Its clients include Wayve, London Medical Imaging, and the AI Centre for Value-Based Healthcare. It also supports AI-based workloads created by independent vendors, such as Cerebras, Blaize, GraphCore, and SambaNova. Additionally, Run:AI supports both CPU and GPU instances on AWS.
Run:AI helps developers and operations teams manage AI infrastructure by reducing the complexity and cost of AI infrastructure. The startup raised a series C round of $75 million from Tiger Global Management and Insight Partners. This round of funding brings its total funding to $118 million. The company plans to use the funds to hire more people and expand globally. It will also consider strategic acquisitions to expand its capabilities.
Atlas platform supports AI workloads on Kubernetes
The Atlas platform orchestrates AI computation and provides GPU access to AI workloads. It utilizes Kubernetes scheduler and software-based fractional GPU technology to create and access multiple GPU nodes. Customers can deploy their own GPU nodes and use the Run AI Atlas platform without changing their code. It also operates on bare metal servers and supports both CUDA and Intel GPUs.
Atlas supports all types of AI workloads, from large-scale training jobs to running production machine learning models. Its modular architecture enables seamless, linear scaling and easier upgrades. Its GPU Abstraction capabilities virtualize GPU resources to optimize infrastructure efficiency and optimize ROI. The Atlas platform supports GPU partitioning, allowing users to pool expensive compute resources to accelerate AI workflows.
Several of the Atlas platform’s ApertureTM platform is used by organizations that help build and operate communities. For example, in Ghana, it can help businesses forecast future infrastructure demand by analyzing data about the local economy and human activity. It can also help businesses determine which areas in Ghana are the most lucrative for new water piping, and can assess the ability of local customers to pay for services.
Runai’s Atlas platform orchestrates AI workloads, including those that use NVIDIA GPUs. It also supports lift & shift and multi-node scaling in the cloud. This portability allows customers to manage their AI workloads across multiple cloud providers with a single control plane.
Samsara IoT platform
Samsara is an industrial Internet of Things (IoT) platform that uses artificial intelligence, cloud computing, and video imagery to transform physical operations. It provides organizations with real-time visibility and analytics to make better decisions and save taxpayer dollars. It combines operational data across multiple departments, enabling public sector agencies to gain actionable insights and improve the day-to-day lives of employees. As a result, it has become the choice of more than 20,000 organizations.
Digibox is an IoT prototyping environment for IoT applications that supports 20 simulated devices and 18 scenarios. It supports a variety of complex devices and scenes, enabling developers to simulate and test IoT applications before deploying them in the real world. Developers can easily customize existing scenes to create new applications or replicate results from scientific research. The platform’s flexible runtime makes it easy to scale the prototype environment.
Samsara is pioneering IoT deployments and was named one of America’s Top AI Companies by Forbes. The Forbes AI 50 lists the top 50 private companies that use artificial intelligence. Samsara also makes use of a unique IoT platform, which dynamically allocates hardware resources for AI training.
Globality raises $100M
AI training startup Globality is raising $100 million in funding to develop and scale its AI training program. It is an end-to-end platform that ensures a level playing field, the best service providers, and the right price. The company is hiring. Globality is based in Israel, but it is growing rapidly in other markets.
AI training startups like Globality are facing a number of challenges. For instance, AI training and development requires an open, collaborative culture between human workers and machines. A key enabler of AI is trust, as machines depend on human data to train their algorithms. Humans will need time to adjust to this paradigm shift, but it is imperative to start building an AI-ready culture early.
Globality has received investment from several tech giants, including Google and Facebook. These companies are placing a large bet on the technology, which could lead to sweeping changes in the global economy. The company also has plans to expand internationally and to create jobs for people with varying degrees of AI skills.
Globality’s technology will improve the way companies train their workers with AI. Using AI training can help employees be more productive. However, it is still important to gain sector-specific expertise. To be successful in this market, AI providers must understand their customers’ value chains and data systems. They must be able to demonstrate benefits at scale.
Run:AI raises $75M
Run:AI provides an open platform to dynamically allocate compute resources for AI training. The company enables companies to abstract the management of AI training infrastructure from data scientists and machine learning (ML) engineers. This allows them to run models with limited GPU memory. The company also enables them to define policies for how their workloads are allocated across a GPU. The company plans to use the money to grow their global teams and to make strategic acquisitions.
Its cloud-based software works with Nvidia GPUs and Intel processors. It is currently used by Wayve, the London Medical Imaging Centre, and the AI Centre for Value Based Healthcare, among others. Run:AI also has partnered with independent vendors like Blaize, Cerebras, GraphCore, and SambaNova.
The startup is focused on helping data scientists run deep learning experiments with less stress. It optimizes training workloads and distributes them across the right hardware. This allows data science teams to run experiments in shorter cycles and stop worrying about computational inefficiencies. This is important for AI training and development.
The company also aims to make AI more accessible to businesses and consumers. By making AI easier to use and more flexible, the company is enabling AI to be more effective than ever. Using AI-powered machine learning, it is possible to automate business processes, which translates into higher revenue and better customer experiences. The company has raised $184M from leading venture capitalists.
