Skip to Content

Managing life science innovation in a world of ever-increasing technological change

Gerard Kerr
20 Jan 2023

The pace of technological change is accelerating due to the law of accelerating returns. Previously emerging technologies are maturing at unprecedented rates and new emerging technologies are entering the spotlight.

Alongside these advances, new technologies are increasingly converging to bring about new solutions that have the potential to rapidly accelerate innovation and disruption over the next decade. We are facing a revolution which will impact every aspect of our lives.

But we have a problem.

While exciting, the acceleration of technological maturation and convergence has a dark side. It is becoming increasingly challenging to manage the growing complexity of solutions underpinning these innovative use cases. In life sciences, it is imperative that organizations carefully manage solutions to maintain high quality patient care while improving safety and lowing risks.

There needs to be a convergence in the way we manage these solutions and provide operations services. Otherwise, each split of responsibilities will introduce more failure points and lead to a muddy view of responsibilities, decreasing stability, increasing time to resolve issues, and hampering change and innovation.

Disparate, siloed teams working on specific functional areas of complex solutions will lead to failure. We have seen evidence of this in the past with the advent of MLOps. One of the key challenges that MLOps sought to solve was the lack of collaboration and ineffective communication between data scientists doing model development, and engineers doing the software development, integration, and operations. MLOps sought to solve this through a practice that brought together the data science, engineering, and operations aspects.

We must learn from lessons of the past to prepare for the future challenges that we will face as more technologies converge. We must coordinate between these disparate practices to form one cross-functional practice.


Enter XOps

XOps is the practice representing the combination of -Ops practices, such as DevOps, MLOps, and DataOps, into one unified framework. These are practices to scale, deploy, maintain & evolve software, models, data engineering pipelines, etc. The reconciliation of these Ops disciplines into XOps combines disparate functions, teams, and processes to manage complex solutions, built upon many different technologies and disciplines. It ensures the continuous delivery of robust, secure, scalable, and trustworthy data-driven AI systems.

For example, think of something like a digital twin of an engine where you have: a data collection/ingestion component, data pipelines, data analytics components, an underlying ontology describing relationships between entities, knowledge graph databases storing the data with its relationships intact, plus edge architecture considerations. You may have some complex visualization components that require UX/AR/VR designers. You also need to include domain experts who understand the context of the data and analysis that needs to be done to it. (Does any of this sound familiar?)

The XOps paradigm pulls together a team with cross-functional knowledge that can work together to build and maintain such a solution. Compared with the traditional method–pulling in 4/5 different teams, each with their own specialization – this is a game changer.

The integration of Ops practices into one unified framework prevents silos from being created among the teams that are managing the solution. Everyone has the high level, end-to-end view of how the solution should be operating. It ensures that, should things go wrong, there is a cross-functional team already in place to provide a quick resolution based on their holistic view. This also means that one can accelerate solution innovation by bringing in a multitude of different viewpoints and technical specialists. The integration of many different lateral perspectives can help reframe problems and apply novel solutions.

In life sciences, this means that XOps can be utilized not only to improve care quality, patient safety, and treatment efficacy – but also to further accelerate healthcare innovation.

Conclusions & Call to Action

How are patients meant to trust increasingly complex solutions built atop converging next-generation technologies and difficult to understand data and analysis methods?
Life science organizations are either starting to or will soon need to reorganize their operations teams. As they deploy more complex solutions to take advantage of the next generation of technologies and innovation, they will find these increasingly difficult to manage. The solution is not to create more and more teams, each focused on limited functional areas. We have seen this before and it doesn’t work. Instead, they must look to XOps.

Author

Gerard Kerr

XOps Manager and Technical Consultant, Capgemini Engineering
Gerard leads our specialist engineering and R&D operations services team in the US. He helps his clients govern, maintain, and evolve their data driven solutions and models. Gerard’s focus is enabling digital transformation, data driven innovation, and protecting his clients’ investments.