Existing software development lifecycle tools can help speed AI/ML products to market
Software developers creating artificial intelligence/machine learning (AI/ML) medical products should move forward with the software development lifecycle (SDLC) models they are already familiar with to bring new products to market, according to experts.Eric Henry, senior quality and regulatory compliance advisor at King & Spalding, and Kip Wolf, president of Thaumazo Bioscience Management, spoke about how to get AI/ML products to fit US Food and Drug Administration (FDA) requirements at the 2023 AI Summit held by the AFDO/RAPS Healthcare Products Collaborative.
They noted that while SDLC models have been around for a long time, software developers often seem stuck on which model to use when developing their product and on getting buy-in from their company executives. Despite the hesitation, they encouraged the audience to trust their abilities and take bold steps to push forward with their product development and not feel the need to reinvent the wheel.
“Don't be cavalier, but if you can find a way to be proactive and maybe a little provocative and find yourself in a schism in your organization where everyone is on hold in analysis paralysis… what we're saying is just do it,” said Wolf. “We're saying it from a position of experience having been there.”
Wolf added that software developers may not feel empowered to take such actions but should feel confident that their quality system management (QMS) gives them the right to use such models to develop AI/ML products.
“If you present a compelling business argument to your leadership that says, ‘Within the existing effective framework that we have that I'm trained in at my organization, I think we can do XYZ, and here are the risks,’ I don't see any leaders standing in the way,” said Wolf.
Henry added that regardless of the SDLC model that developers choose, they should keep in mind people, processes and technology when developing their AI/ML software.
“Keep in mind, whether you're developing in weeks or months or years, that we should always engage the people,” he said. “Make sure you got a good stakeholder map.”
From experience, Henry said there’s typically a gap between the developers and stakeholders, and it’s best not to wait till the end when the software has been tested and features have been added before asking whether all the stakeholder needs have been met. He noted that AI/ML vernacular has become much more common these days which is creating more push within companies to develop such products faster.
“In the past, we had the luxury of developers to kind of build something and have a life cycle and plan when the release strategy would happen,” said Henry. “[Now] we're kind of chasing a sense of urgency where I think it's incumbent to minimize the risk of failing to meet the value proposition at the end to apply the agile development methodology to stakeholder management along the way as well.”
Henry also noted that he moved into the medical device software space in the early 2000s from working in other industries that use SDLC models and found that the medtech industry was five years to a decade behind in terms of adopting process maturity and SDLC models.
×
Welcome to the new RAPS Digital Experience
We have completed our migration to a new platform and are pleased to introduce the updated site.
What to expect: If you have an existing login, please RESET YOUR PASSWORD before signing in. After you log in for the first time, you will be prompted to confirm your profile preferences, which will be used to personalize content.
We encourage you to explore the new website and visit your updated My RAPS page. If you need assistance, please review our FAQ page.
We welcome your feedback. Please let us know how we can continue to improve your experience.