MILS Group is a startup founded in 2015 out of passion to bring innovation to Life Sciences and Healthcare by harnessing the growing effectiveness of Artificial Intelligence and Machine Learning.

MILS Group has launched a research project that will focus on finding an effective low-cost solution for the integration of Artificial Intelligence and Machine Learning into process development and control for upstream/downstream biologic drug substance manufacturing.

Our current project’s objective is to find an optimal pathway for application of Machine Learning in development and manufacturing of biologic drugs. We will focus on the cost saving and quality enhancing potential of AI/ML solutions in up- and down-stream bio-manufacturing process development and control to:

Decrease cost of manufacturing process:

• Reduce process variability to avoid deviation from CQA and batch loss
• Improve dynamic modeling of process variables alterations without jeopardizing validation requirements
• Provide real-time Bayesian modeling of off-line analytical tests and postproduction assays
• Collect data for effective adoption of QbD and PAT

Increase speed and accuracy of process development:

• Predict complex interactions of multiple cross-correlated process variables
• Reconcile term indicators and state conversion variables, automate data recalibration for batch-to-batch/process-to-process comparison
• Scale up/down with non-linear change of process variables
• Automate knowledge generation for multi-facility data collection, aggregation and alignment

To facilitate adoption, the AI solutions for bio-manufacturing process development and control should deliver a substantial value to bio-processing scientists:

Increase scientific productivity and rate of success:

• Automate time-consuming, repetitive data-driven process development tasks and decision making
• Speed up experimentation with effective scale up process modeling

Reduce complexity of everyday process development and control workflow

• Decrease ambiguity of design space with complex interaction between process variables and CQA
• Increase accuracy of predictive modeling to enhance dynamic process control

Create higher job satisfaction and exciting professional development perspective:

• Allow to focus on exciting experimentation without fear of jeopardizing validated process
• Reduce stress related to experimentation failure and material loss
• Support career transition into new PAT/AI driven era

Some initial project findings are summarized in the article:

Artificial Intelligence in Biopharmaceutical Manufacturing

More publications to follow.

We would love to get in touch with you!