The world’s inhabitants is increasing at a fast fee. The rising international inhabitants requires modern options to provide meals, fiber, and gasoline, whereas restoring pure assets like soil and water and addressing local weather change. Bayer Crop Science estimates farmers want to extend crop manufacturing by 50% by 2050 to fulfill these calls for. To assist their mission, Bayer Crop Science is collaborating with farmers and companions to advertise and scale regenerative agriculture—a future the place farming can produce extra whereas restoring the surroundings.
Regenerative agriculture is a sustainable farming philosophy that goals to enhance soil well being by incorporating nature to create wholesome ecosystems. It’s based mostly on the concept agriculture ought to restore degraded soils and reverse degradation, fairly than maintain present situations. The Crop Science Division at Bayer believes regenerative agriculture is foundational to the way forward for farming. Their imaginative and prescient is to provide 50% extra meals by restoring nature and scaling regenerative agriculture. To make this mission a actuality, Bayer Crop Science is driving mannequin coaching with Amazon SageMaker and accelerating code documentation with Amazon Q.
On this submit, we present how Bayer Crop Science manages large-scale knowledge science operations by coaching fashions for his or her knowledge analytics wants and sustaining high-quality code documentation to assist builders. Via these options, Bayer Crop Science tasks as much as a 70% discount in developer onboarding time and as much as a 30% enchancment in developer productiveness.
Challenges
Bayer Crop Science confronted the problem of scaling genomic predictive modeling to extend its pace to market. It additionally wanted knowledge scientists to give attention to constructing the high-value basis fashions (FMs), fairly than worrying about developing and engineering the answer itself. Previous to constructing their answer, the Resolution Science Ecosystem, provisioning an information science surroundings may take days for an information staff inside Bayer Crop Science.
Resolution overview
Bayer Crop Science’s Resolution Science Ecosystem (DSE) is a next-generation machine studying operations (MLOps) answer constructed on AWS to speed up data-driven determination making for knowledge science groups at scale throughout the group. AWS providers help Bayer Crop Science in making a related decision-making system accessible to 1000’s of knowledge scientists. The corporate is utilizing the answer for generative AI, product pipeline developments, geospatial imagery analytics of discipline knowledge, and large-scale genomic predictive modeling that may enable Bayer Crop Science to turn out to be extra data-driven and improve pace to market. This answer helps the information scientist at each step, from ideation to mannequin output, together with the complete enterprise determination file made utilizing DSE. Different divisions inside Bayer are additionally starting to construct the same answer on AWS based mostly on the success of DSE.
Bayer Crop Science groups’ DSE integrates cohesively with SageMaker, a totally managed service that lets knowledge scientists shortly construct, prepare, and deploy machine studying (ML) fashions for various use circumstances to allow them to make data-informed selections shortly. This boosts collaboration inside Bayer Crop Science throughout product provide, R&D, and business. Their knowledge science technique not wants self-service knowledge engineering, however fairly gives an efficient useful resource to drive quick knowledge engineering at scale. Bayer Crop Science selected SageMaker as a result of it gives a single cohesive expertise the place knowledge scientists can give attention to constructing high-value fashions, with out having to fret about developing and engineering the useful resource itself. With the assistance of AWS providers, cross-functional groups can align shortly to scale back operational prices by minimizing redundancy, addressing bugs early and infrequently, and shortly figuring out points in automated workflows. The DSE answer makes use of SageMaker, Amazon Elastic Kubernetes Service (Amazon EKS), AWS Lambda, and Amazon Easy Storage Service (Amazon S3) to speed up innovation at Bayer Crop Science and to create a custom-made, seamless, end-to-end person expertise.
The next diagram illustrates the DSE structure.
Resolution walkthrough
Bayer Crop Science had two key challenges in managing large-scale knowledge science operations: sustaining high-quality code documentation and optimizing current documentation throughout a number of repositories. With Amazon Q, Bayer Crop Science tackled each challenges, which empowered them to onboard builders extra quickly and enhance developer productiveness.
The corporate’s first use case centered on mechanically creating high-quality code documentation. When a developer pushes code to a GitHub repository, a webhook—a light-weight, event-driven communication that mechanically sends knowledge between functions utilizing HTTP—triggers a Lambda perform via Amazon API Gateway. This perform then makes use of Amazon Q to research the code modifications and generate complete documentation and alter summaries. The up to date documentation is then saved in Amazon S3. The identical Lambda perform additionally creates a pull request with the AI-generated abstract of code modifications. To keep up safety and adaptability, Bayer Crop Science makes use of Parameter Retailer, a functionality of AWS Programs Supervisor, to handle prompts for Amazon Q, permitting for fast updates with out redeployment, and AWS Secrets and techniques Supervisor to securely deal with repository tokens.
This automation considerably reduces the time builders spend creating documentation and pull request descriptions. The generated documentation can be ingested into Amazon Q, so builders can shortly reply questions they’ve a few repository and onboard onto tasks.
The second use case addresses the problem of sustaining and bettering current code documentation high quality. An AWS Batch job, triggered by Amazon EventBridge, processes the code repository. Amazon Q generates new documentation for every code file, which is then listed together with the supply code. The system additionally generates high-level documentation for every module or performance and compares the AI-generated documentation with current human-written documentation. This course of makes it doable for Bayer Crop Science to systematically consider and improve their documentation high quality over time.
To enhance search capabilities, Bayer Crop Science added repository names as customized attributes within the Amazon Q index and prefixed them to listed content material. This enhancement improved the accuracy and relevance of documentation searches. The event staff additionally applied methods to deal with API throttling and variability in AI responses, sustaining robustness in manufacturing environments. Bayer Crop Science is contemplating growing a administration aircraft to streamline the addition of latest repositories and centralize the administration of settings, tokens, and prompts. This is able to additional improve the scalability and ease of use of the system.
Organizations seeking to replicate Bayer Crop Science’s success can implement comparable webhook-triggered documentation era, use Amazon Q Enterprise for each producing and evaluating documentation high quality, and combine the answer with current model management and code assessment processes. Through the use of AWS providers like Lambda, Amazon S3, and Programs Supervisor, corporations can create a scalable and manageable structure for his or her documentation wants. Amazon Q Developer additionally helps organizations additional speed up their improvement timelines by offering real-time code ideas and a built-in next-generation chat expertise.
“One of many classes we’ve discovered over the past 10 years is that we need to write much less code. We need to focus our time and funding on solely the issues that present differentiated worth to Bayer, and we need to leverage every part we will that AWS gives out of the field. A part of our purpose is lowering the event cycles required to transition a mannequin from proof-of-concept section, to manufacturing, and in the end enterprise adoption. That’s the place the worth is.”
– Will McQueen, VP, Head of CS World Knowledge Property and Analytics at Bayer Crop Science.
Abstract
Bayer Crop Science’s method aligns with trendy MLOps practices, enabling knowledge science groups to focus extra on high-value modeling duties fairly than time-consuming documentation processes and infrastructure administration. By adopting these practices, organizations can considerably cut back the effort and time required for code documentation whereas bettering general code high quality and staff collaboration.
Be taught extra about Bayer Crop Science’s generative AI journey, and uncover how Bayer Crop Science is redesigning sustainable practices via cutting-edge know-how.
About Bayer
Bayer is a world enterprise with core competencies within the life science fields of well being care and diet. Consistent with its mission, “Well being for all, Starvation for none,” the corporate’s services and products are designed to assist folks and the planet thrive by supporting efforts to grasp the most important challenges offered by a rising and growing old international inhabitants. Bayer is dedicated to driving sustainable improvement and producing a optimistic influence with its companies. On the similar time, Bayer goals to extend its incomes energy and create worth via innovation and development. The Bayer model stands for belief, reliability, and high quality all through the world. In fiscal 2023, the Group employed round 100,000 folks and had gross sales of 47.6 billion euros. R&D bills earlier than particular objects amounted to five.8 billion euros. For extra info, go to www.bayer.com.
Concerning the authors
Lance Smith is a Senior Options Architect and a part of the World Healthcare and Life Sciences business division at AWS. He has spent the final 2 a long time serving to life sciences corporations apply know-how in pursuit of their missions to assist sufferers. Exterior of labor, he loves touring, backpacking, and spending time together with his household.
Kenton Blacutt is an AI Advisor inside the Amazon Q Buyer Success staff. He works hands-on with clients, serving to them resolve real-world enterprise issues with cutting-edge AWS applied sciences. In his free time, he likes to journey and run an occasional marathon.
Karthik Prabhakar is a Senior Purposes Architect inside the AWS Skilled Companies staff. On this function, he collaborates with clients to design and implement cutting-edge options for his or her mission-critical enterprise programs, specializing in areas equivalent to scalability, reliability, and price optimization in digital transformation and modernization tasks.
Jake Malmad is a Senior DevOps Advisor inside the AWS Skilled Companies staff, specializing in infrastructure as code, safety, containers, and orchestration. As a DevOps marketing consultant, he makes use of this experience to collaboratively works with clients, architecting and implementing options for automation, scalability, reliability, and safety throughout all kinds of cloud adoption and transformation engagements.
Nicole Brown is a Senior Engagement Supervisor inside the AWS Skilled Companies staff based mostly in Minneapolis, MN. With over 10 years {of professional} expertise, she has led multidisciplinary, international groups throughout the healthcare and life sciences industries. She can be a supporter of ladies in tech and at present holds a board place inside the Girls at World Companies affinity group.
