Increase packaging efficiency with AWS SageMaker for a pharmaceutical company Increase packaging efficiency with AWS SageMaker for a pharmaceutical company

Customer Background

Headquartered in Dallas, our client is a pharmaceutical company established in 1965. The client manufactures a variety of drugs every day. As per the pharmaceutical industry norms, the company must follow strict packaging and quality assurance guidelines. Our client’s goal was to reduce the drug rejection rate due to problems related to damaged containers, missing tablets in blister packs and packaging of broken tablets.

  • Industry

    Pharmaceutical

  • Technologies / Platforms / Frameworks

    Amazon SageMaker, Amazon Rekognition, AWS Lambda

Challenges

Packaging and quality assurance are a few of the most challenging environments in the pharmaceutical industry. The industry has strict guidelines for packaging and quality assurance.

Our client’s primary concern was a high drug rejection rate due to mishaps in packaging and quality assurance. Further, the scene of rejected drugs also increases the overhead costs as they are checked during shipping. Hence, assessing the packages while still in the production cycle is crucial.

The client allocated dedicated personnel in packaging processes to identify defective packs or containers at different stages of the production cycle. This is, of course, prone to error and depends on an individual’s expertise.

To overcome these problems, the client established an upgraded setup. The client started using cameras to capture visual information but faced difficulties extracting valuable insights. Overall, the disasters in packaging processes were affecting our client’s:

  • brand reputation
  • customer satisfaction
  • profits

Solutions

While solving the client’s problem, our focus was not only on quality check automation but on enhancing accuracy. Hence, we combined artificial intelligence with visual analysis. We trained a model to understand things exactly as humans learn from visual inspection. This is highly important to nullify the packaging processes and quality assurance inefficiencies.

After analyzing our client’s needs in the core areas where automation and machine learning could help, we leveraged Amazon SageMaker. Using the platform, we built an ML model from the visual data that accurately detect inadequacies in packaging processes.

Overall, our AWS experts designed a custom machine learning model based on the camera feed of the packaging processes. Our ML experts trained the model with different images of drugs and packaging variants. As the output, the model analyzes every package faster and accurately, assigning a label based on the quality of manufactured drugs and the packaging.

The sorting machine receives the feed from the model’s analysis. Based on this, it rejects the container of defective packaging or drugs. Over time, the client can use the data to create a model that can explain the reasons behind faulty packaging and compromises in quality assurance.

Benefits

  • Increased defect detection rate
  • Identifying packaging issues before shipping
  • Reduced drug rejection rate boosts sales
  • Optimized labor costs
  • Simplified quality assurance system
  • Improved brand reputation and customer satisfaction
AWS Sagemaker

Need more information?

Tell us what you are looking for and we will get back to you right away!