Nyun Ai Private Limited

In their own words

Deep learning has become a go-to-solution for several problems of science and engineering disciplines, be it for the increased productivity in mines or even early detection of dangerous medical conditions, such as COVID-19. There are hundreds or even thousands of real-world applications where deep learning solutions have been deployed in the real-world. It is speculated that till now, not even 1% of the actual potential of deep learning in the real-world has been explored, and we will see more applications of deep learning in the coming years. Among various current challenges that exist, one important problem related to deep learning that needs to be addressed is to make them computationally efficient. Various ways exist to make models computationally efficient. Among these, one of the popular research directions is to compress the neural network architectures with minimal to no modifications in the performance. On a technical level, this can be achieved through quantization of the network, binarization of its parts, structured and unstructured pruning. While these approaches have been proved valuable, their commercial value is still to be exploited. Nyun AI aims towards becoming the pioneer in designing extremely efficient neural networks. We will be releasing several products namely such as Nyun API and Nyun Store. Nyun API: This will include some of the popular pruning/quantization/binarization algorithms in a modular manner. Further, it would be designed in a manner that end-users can easily develop efficient models based on the functionalities of the package. Note that while the package itself can make model compression relatively easy, pruning and quantization are still considered among highly-skilled tasks, and the end-user is expected to approach us in future for the development of these compressed models. Nyun Store: Nyun store will have a suite of products that can solve practical daily problems. One of the products is Object Detection: Our product once installed by the end user can count the person entering/exiting a door. This will be easy to set up and will not require any coding from the end user side. Nyun AI also provides services to enterprise clients looking to develop efficient deep learning models. We target 2 main sectors which are edge computing and computer vision. The deep learning models developed by us are reliable, have less inference time, consume less data, require less processing power and are cost efficient. As stated earlier these efficient models have a wide variety of applications. Few of them are mentioned below. 1. Manufacturing and Natural Resources • Predictive maintenance of machines • Remote quality inspection with high resolution image • Collaborative robots 2. Smart Cities • Smart video surveillance • Smart home/building • Smart traffic and mobility management • VR/AR in education and healthcare 3. Energy and Utilities • Distributed energy resource management • Smart grids • Grid backhaul and backbone • Grid access communication • Remote site monitoring 4. Healthcare • Mobile medical data sets delivery in real time • Cloud robotics for assisted living or rehabilitation • Remote health monitoring • Precision tele-robotics surgeries 5. Agriculture • AR support for eLearning and expert advice in remote areas • Predictive maintenance of farming equipment using sensor data • Livestock tracking 6. Retail • Optimal fusion of the virtual and physical worlds for a personalized shopping experience • Predicated preference based on customer location 7. Automotive, Transportation and Logistics • High-speed connectivity with real-time high-resolution vehicle video surveillance • Connected traffic management through the cloud to assist operational decision-making • Autonomous cars

Written by the company on its public Startup India profile. Not verified by us.

Is this your company?

Everything here comes from public government records. If something is wrong, the authoritative place to correct it is the Startup India portal itself, and we will pick the correction up on the next monthly snapshot. To flag an error on our side, use the report link or write to trust@indianstartupmap.com.

How we built this page Where the data comes from How you may reuse it