It is very important to have a good grasp of the concepts. But mere concepts are not enough. In fact they are meaningless until we have working code that does the job. Here, is an introduction to the requisite software skills.
The software implementation requires mastery of the programming language and the libraries therein. We do have some tools (e.g. EZDL) that can help us perform the tasks without a single line of code. Yet these have limitations. Also, for using such tools effectively, we need to understand the core concepts of the code that runs underneath.
AI development had stagnated for many years due to the lack of processing power and data. The amount of processing power required for building meaningful AI models is massive and it makes very little sense for anyone to acquire it on campus. It is far more sensible to do such processing on cloud.
The cloud providers today provide a lot more than just processing power. AWS, Google Cloud, Azure, Digital Ocean, RedHat, Alibaba, IBM and most popular companies help their customers with most of the system required for training and deploying models. Let us have a look at the offerings of the major players
Apart from these, we have many other cloud service providers that focus specifically on AI products.
Machine Learning and AI in general; are too vast to grasp in a few days. The domain is developing and growing every day. Each day comes up with a new set of surprises and each day showers newer developments. These blogs continue to grow as I learn the subject.