Artificial intelligence (AI) and Machine Learning (ML) are two technologies that hold vast potential for enterprises, which is why they have quickly become front of mind for many members of the boardroom.
Machine learning in particular is being touted for its ability to help organisations strengthen their competitive advantage and drive innovation. As such decision makers are already starting to steer their respective companies towards an ML transformation.
With success or failure in this regard meaning the difference between disruptive change or being left behind, it begs the question of whether these executives are steering their company in the right way.
The answer for tapping into data
According to Amazon Web Services‘ senior data analyst, Yotam Yarden, “there is literally no industry segment which can’t leverage data to improve and create new business models.”
“Meanwhile, data has never been easier and less expensive to collect, store, analyse, and share. Many enterprises are building their data lakes today precisely for this reason. But, is your organisation taking full advantage of its data?,” he posits.
Machine learning is a technology that can help organisations make the most of their data, but there is still a lack of knowledge on how to implement it effectively.
In order to gain a better understanding of Machine Learning, as well as how executives need to approach it, Yarden unpacks seven significant aspects for them to know.
1. Be business driven and customer focused
In order to address your organisation’s biggest challenges, Yarden advocates for looking at a focused business challenge and working backwards towards a solution.
“Too many companies try to apply “self-driving cars” or “genome-sequencing” algorithms to a sales funnel optimisation challenge just because they hired an expert in this field, while often there are models that better fit the task and bring higher value at lower costs,” notes the analyst.
“Don’t keep your data science team in the IT department alone. Rather, giving ownership of the data science team to a business stakeholder can invigorate your organisation, and unlock new revenue streams and tremendous cost savings,” he advises.
2. Iterate fast and simple
When it comes to bringing your ML system into production, it’s often best to do so in smaller iterations, says Yarden.
“Conducting small iterations through tests, proof of concepts and pilots will help your team to bring ML workloads into production faster, and in a higher quality,” he explains.
Yarden also notes that even if your organisation is not utilising the latest ML technologies, adopting a quick iteration approach holds greater value than a long development process.
“Only by experimentation, experience, and adaptation, can you realise the full potential of your ML product. Fail fast and improve often,” the analyst says.
3. To centralise or decentralise
When developing your organisation’s ML system, there will come a choice to centralise or decentralise your team. For Yarden, it’s not simply a choice of one or the other, but should be driven by flexibility.
“ML applications, like any other piece of software, require maintenance, updates, and support,” adds Yarden.
“A centralised team may be effective at low-scale, but once you start expanding, innovation might suffer. Imagine a large innovation team who is working on multiple innovative projects, it is inevitable that at some point a substantial portion of the team’s work would be operating ongoing projects,” he says.
4. Tackling big roadblocks
In the ML transformation journey, your data scientists and developers will invariably encounter roadblocks. Yarden says many of these are quite easily overcome though, with proper planning ahead of time often being the solution.
Should there be a lack of sufficient data for example, he advises collecting data well before a project kick-off gets underway. Another common issue is data being inaccessible, which can be handled by ensuring that relevant data samples are obtained.
5. Data science and DevOps are no longer divided
In AWS’ experience, customers often have ML modellers create specifications for developers to implement. In order to tackle this, there are a wide range of tools in place that allow data scientists to be involved in the engineering, and vice versa, according to Yarden.
“Thankfully, technology is evolving at an increasing pace and new tools are continually released. It has never been easier for experts to expand their capabilities and cross over into new domains,” he continues.
6. Maintain the ratio
One of the other big questions that AWS has to contend with is the number of data scientists and engineers that should be onboard for a project, especially as it pertains to the ratio thereof.
Yarden says that ratio is dependent on the maturity of a project, as well as how much data there is to analyse. More data means more scientists will be needed, and less data necessitates the need for a larger number of engineers.
“As a rule of thumb, plan to have 2-3 engineers for every data scientist in the building phase, and 1:1 when a system is already deployed,” he advises.
7. Make KPIs clear
KPIs (Key Performance Indicators) are a great way of ensuring that ML projects have regular checkpoints in place to ensure things are running smoothly and according to plan.
It not only helps in defining whether an objective is overly ambiguous, or if a goal has indeed been achieved.
“Each measure could lead to a different recommendation system. Having clear goals & KPIs will help you plan and execute more effectively,” says Yarden.[Image – CC 0 Pixabay]