HALO graphic

Machine Learning at your Fingertips

Holistic Artificial Life Outcomes (HALO) Compute is a heterogeneous computing platform capable of operating across CPUs, GPUs, FPGA, and APUs.

HALO makes machine learning accessible. Whether you are executing neural networks or processing Map Reduce ETL jobs, HALO offers ease of use and performance. Through HALO, algorithms are easily orchestrated, allowing for complex models and simulations. You can choose to bring your own algorithm or use algorithms from our extensive cache to jumpstart any project.

Designed for Performance

HALO utilizes specialized native drivers to accelerate algorithm processing that scales with the volume of data. It is capable of decomposing highly iterative algorithms across disparate platforms resulting in a highly-distributed environment. Our platform treats all compute units like a cloud, allowing dynamic allocation to any running algorithm to scale up or down on demand. HALO does not just learn from data, it learns in real-time from the analysts that use the platform. User actions and decisions are assimilated as data.

HALO is specifically designed to execute machine learning tasks while disconnected from the data center. Our distributed capability results in a swarm of coordinated activities resilient to single or multiple points of failure while in real-time operation. HALO brings the AI supercomputer as a portable swarm to reality. Our Decision Support UI Accelerators and Open Architecture RESTful services make outcomes presentation and integration very easy.
halo table
HALO is a diverse computing environment and can extend the command center to the field in a secure distributed fashion. If you would like to learn more about HALO or request a demo, contact Mr. Jared Spigner or Dr. Ben Nguyen at 703.890.8500, or visit our Contact Us page.