Low-Code AI Platform

About

There is an unfilled gap between AI research entities and the actual deployment of AI by the industry, together with a shortage of available AI talent makes real world AI adoption slow. Critical information slips through the cracks with inhouse talent spending most of their time sifting through information manually. Whole industries need to put in place intelligent automations that not only make sense of this big data but also learn on the go while responding or taking actions. It is not only about getting knowledge from data, but as well automating actions with this knowledge to make businesses agile and resilient. Our AUTOMAISE low-code AI Platform creates enterprise grade intelligent automations. With a unique mix of Deep Learning, NLP, Big Data, and bits of RPA we enable hyper-automation delivering automation rates above 80%. We are on a mission to democratize enterprise grade AI automation to everyone. The AUTOMAISE Platform is equipped with state-of-the-art AI and Machine Learning (ML) algorithms, from simple logistic regressions to the highly complex Deep Learning architectures, such as LSTM, Convolutional Neural Networks and Transformer based models. The range of tasks that is possible to tackle with our Platform ranges from classical classification tasks to Computer Vision tasks. At the core of our platform is a state-of-the-art Automated Machine Learning (autoML) layer that is able to adapt, combine and automate our library of ML algorithms, including all the necessary data wrangling (specifically, data cleaning, pre-processing and feature engineering) that is required so that such algorithms can unveil their full potential. For this challenge, this Automated Machine Learning layer will be the backbone of our approach, in particular, the autoML Computer Vision module that we developed. This component enables a quick and robust search for a ML pipeline that gets the most out of your image database, with a particular emphasis on classification tasks. There are two major advantages of using such automated approach: 1) very strong baseline models achieved in a very short period of time. Our autoML layer is embedded with extended expertise both in terms of ML algorithms and data wrangling, and since its always being improved, is able to accumulate knowledge from one project to another 2) faster model iterations for the data scientists. Since we reduce the time that they spent writing tedious and error prone code for tasks such as data cleaning or feature engineering, our data scientists can focus on testing hypothesis (specifically, new features or new learning algorithms) to evolve the model even further.