AIron

We decided to change the brand to AIron, we had new logo, web, etc.

Also, we focused on MLOps (Machine Learning Operations). MLOps is the procedure of creating repeatable and reproducible ML pipelines to improve control over data preparation, training and scoring processes.

We also focused our services on the three biggest clouds: Azure, AWS and Google Cloud.

We started getting more and more projects, initially in Europe, but then we started to expand to the US.

We did projects in different ML areas: object detection, fraud detection, image classication, dimensionality reduction and many other areas.

We were growing and evolving fast.

Our first customer was in the area of text classification. We used convolutional neural networks (also named CharCNN model) and created an API that would classify a sentence into several classes.

Back then using CNN for text was a new thing, these days we are using other techniques.

Back then, not that many companies were using machine learning. However, the field was starting to attract attention. Particularly, the area of deep learnig started to make dramatic advances in AI, so we turned our attention to this area.

The adventure has started.