Over 20 years of experience in development of compiler back-ends for high-performance chip and system vendors, as well as application tuning for GPUs and multicore processors (CPUs).  Over 10 years of experience in computational statistics and workflow acceleration for scientific applications.

The development of predictive analytics and machine-learning solutions requires highly-specialized skills.  The developer is often more focused on a particular problem domain or technique, and is less concerned with broader concerns such as extensibility and maintainability.  While in-house talent may be able to address some of these needs, performance is often neglected.  As the size of supported data sets grows, though, performance can suddenly seize attention.

The role of InPredict is to smooth the transition from a working prototype or in-house solution to a scalable, high-performance application.  We offer refactoring and rescaling services, as well as consultation, to facilitate your redeployment.  We specialize in parallelization, whether across a cloud implementation or within a GPU.  We can also tailor your application to run in embedded environments and SOCs.