4 Artificial Intelligence (AI) skills IT pros must have

As Artificial Intelligence (AI) technologies become more mainstream in the enterprise, what skills set you apart? Individuals and IT teams, focus on these four
228 readers like this.
AI artificial intelligence

Artificial Intelligence (AI) has arguably become a household term in modern enterprises. By now, most companies have embraced some type of business initiative that includes AI in their digital transformation.

Artificial Intelligence is a broad term, but much current research and development focuses on machine learning (ML), a subdiscipline whereby machines learn from data as opposed to being explicitly programmed.

AI skills to watch

With AI and ML targeting a broad spectrum of enterprise users, IT professionals must develop new skills to succeed in this emerging space. Here are four examples.

1. Framing business problems in the context of data

An understanding of the business and its most pressing problems is a transcendent competency for any IT professional. However, AI-driven projects require solutions to be framed and rationalized in the context of data that is directly or indirectly available to the enterprise.

[ Want best practices for AI workloads? Get the eBook: Top considerations for building a production-ready AI/ML environment. ]

The essential question is whether such data has the potential to solve the business problem at hand. While the answer is not always immediately obvious, it begins with a hypothesis stemming from prior analysis or perhaps simply based on intuition. For example, a business experiencing high customer churn might hypothesize that recent changes in commercial activity could predict future attrition.

2. Data engineering

Most enterprises have an abundance of data, but leveraging it for AI/ML projects can be challenging. Preparing data for analysis and machine learning is usually the long pole in the tent because most organizations don’t understand the investment required during this phase.

Increasingly, enterprises are realizing the importance of building systems and processes that automate the acquisition, transformation, and delivery of data to organizations involved in analytics and AI/ML projects. These enterprises understand that data should be a first-class asset on par with code and that the core principles of software engineering should be applied in a similar fashion.

While all IT professionals should have basic data transformation skills, we will likely see the emergence of centralized data engineering teams whose primary purpose is to develop and deploy automated data pipelines that deliver high-quality data at scale.

[ Read also: 6 misconceptions about AIOps, explained. ]

3. Toolchains and languages for machine learning

The tools and infrastructure for machine learning have evolved radically over the past decade, in both open source and commercial offerings. Access to cutting-edge technologies once reserved for only the most elite researchers and practitioners has been democratized, with fully integrated toolchains and services from all major cloud providers.

Various programming languages are used for machine learning, but Python is the most common. Much of its success is due to an active and vibrant community as well as the availability of libraries that implement virtually all the popular algorithms. 

A project that might once have required a data scientist may now be done by IT professionals.

The differentiation of skills between data scientists and software engineers has blurred in recent years due to advances and accessibility in tooling. A project that might once have required a data scientist may now be done by IT professionals.

4. Evaluating model performance

The technology for model selection and training, particularly with integrated toolchains from popular cloud providers, is evolving to a point where certain laborious decisions often made by data scientists are now being done automatically in software. A clear example is the selection of a model that yields the best performance while also generalizing well.

[ Check out our primer on 10 key artificial intelligence terms for IT and business leaders: Cheat sheet: AI glossary. ]

Even as these tools become more advanced, IT professionals should have a general understanding of machine learning concepts – particularly in evaluating model performance and correlating feature selection with predictive quality.

Leveraging artificial intelligence and machine learning to improve business outcomes is quickly becoming table stakes for modern enterprises as they navigate digital transformation initiatives. Embracing these evolving technologies requires IT organizations to develop new skills aimed at using data to solve business problems. To better enable organizations engaged in AI projects, IT teams should also implement new systems and processes that automate the acquisition, transformation, and delivery of data.

A variety of resources are available online to help IT professionals gain the AI and ML skills they need. Coursera.org offers an excellent introductory course that teaches the fundamentals of machine learning. Additionally, all major cloud providers, including AWSAzure, and Google, offer training for their AI services and integrated toolchains. While many of these online courses are free, some – such as certification programs – involve a fee.

[ Culture change is the hardest part of digital transformation. Get the digital transformation eBook: Teaching an elephant to dance. ]

David Edwards
David Edwards is the Chief Technology Officer at Vendavo. David brings an extensive and distinguished background in software development and engineering leadership to Vendavo, particularly in the areas of application and data architectures, scalable computing and distributed systems.