Infrastructure for AI
Successful AI & ML projects
A key point for successful machine learning and artificial intelligence projects is a fast and iterative proceeding for evaluation of structures and hyperparameter as pointed out in the Book Machine Learning Yearning by Andrew NG available at https://www.deeplearning.ai/programs/ .
To support such approaches high-performance hardware and easy-to-use software stacks are required to support the complete workflow from data ingestion to application of learned models wherein machine learning itself comprises only a small fraction, see in paper Hidden Technical Debt in Machine Learning Systems (https://papers.nips.cc/paper/2015/file/86df7dcfd896fcaf2674f757a2463eba-Paper.pdf)
If you require support in selection of hardware ranging from entry-level workstations for AI up to scaleable cluster systems as shown in the figure above do not hesitate to contact me.
As an enthusiast for open-source software I further provide consulting for software frameworks that scale from single workstations to large cluster systems for example based on Kubeflow