Explore the selection of our case studies in the areas of Artificial Intelligence and HPC. To find out more, contact us directly.
In every project we deliver, we combine some of the coolest and cutting edge machine and deep learning, optimization and acceleration techniques to solve real-world problems and meet our client’s needs. The philosophy of the company has been built around the model where we strive to benefit from the latest & greatest research results in the universities while injecting these straight into the business world. Thus we have PhD+ experts on our team who continue researches seeking solutions to emerging problems and constantly work on finding new applications of their inventions.
The unique combination of world-class researchers and highly-qualified, fast-learning engineering teams helps us bring true innovations to the market. Thru innovations we help our clients explore concepts that weren’t previously possible.
- byteLAKE’s Starter Kit: Edge AI for Traffic Analytics
- byteLAKE’s Starter Kit: Edge AI for Retail
- AI inferencing on Lenovo’s edge devices (NVIDIA GPU, Intel Movidius)
- Federated Learning: collaborative AI models training in distributed IoT environment
- Edge AI device prototype for Sound Analytics (new algorithm design and adaptation, edge device prototyping, data collection and processing – whole cycle)
- Computer Vision powered quality inspection for Factory 4.0
Machine & Deep Learning
- Forestry Management with AI
- AI automation for workflow system and RPAs
- Drones footage analytics and illegal dumping areas detection
- AI engine to accelerate and automate documents processing
- Deep Neural Network optimized for NVIDIA GPU
- Time series data analytics with AI
- Face analytics with Azure Cognitive Services
- Transfer learning to improve automatic tagging of images
- Intelligent Restaurant / Hotel (new product incubation, ongoing)
- CFD Kernels adaptation and optimization to CPU-only and CPU+Xilinx Alveo FPGA architectures
- Weather simulation engine (CFD/MPDATA) acceleration with GPU (Piz Daint)
- HPC application ported and optimized to CPU+Nvidia architecture (CUDA)
- Research: finding a way to advance machine learning techniques for HPC optimizations
- Upgraded SLURM to increase nodes utilization and increase the performance of jobs set
- Crypto Mining Algorithms adaptation and acceleration with Xilinx Alveo FPGA (research ongoing)
- AI accelerators benchmark: NVIDIA GPU vs. Intel Movidius
- Artificial Intelligence deployments guideline: on-device or in the cloud or data center?
- AI edge devices benchmark: powered by NVIDIA Quadro P1000 vs. AMD industry customized working sample GfX
- Series of benchmarks of various AI workloads for different HPC architectures and storage impact analysis
(more storage intensive benchmarks on the roadmap)
As a team, we do amazing things.