Agriculture

Speeding Up Environmental Analytics with MLOps

An environmental analytics project had been stuck in proof-of-concept since 2017, slow and able to handle only one query at a time. We moved it to AWS and redesigned how its key calculations worked, delivering ahead of schedule.

Syngenta

2 min read
4years
the environmental querying project sat in POC before we joined in 2021
1query
at a time was all it could handle when we found it

About the client

Syngenta is a global provider of agricultural science and technology. They specialize in solutions that transform how crops are grown and protected for the benefit of farmers, society, and the planet.

Why were we there?

Syngenta is heavily investing in R&D in order to be the global leader in agricultural technology. They have a large team of data specialists working on innovative solutions all the time, and needed a data science platform on AWS that empowers specialists and increases their productivity.

One of the data science initiatives is a high-profile project concerned with querying environmental features from around the world that started in 2017. Due to its complex nature, the project was stuck in a POC phase — execution times were slow, and the product could only handle one query simultaneously.

How we helped

As experts in data science on AWS, we helped Syngenta by designing their data science platform in accordance with the best practices of doing data science projects at scale. We introduced MLOps to streamline a data scientist's development flow, and rolling out our tagging strategy helped with access and cost management — including a Cost Overview dashboard built on those same tags.

We joined the environmental querying project in 2021, moved it to AWS, and redesigned how certain key elements were calculated. We delivered the project faster than expected, which led to new data science initiatives in modeling, predictions, and optimization.

The impact we made

  • Faster and more convenient onboarding of data scientists from different teams

  • Clear insight into costs

Technologies we used

AWSAmazon SageMakerMicroservicesMachine LearningData SciencePythonSQL
“The team is highly skilled, responsive, and a pleasure to work with, including their ability to create a trusted relationship with business teams. Partnering with them has been a game-changer for our business!”
Aart LabeeGlobal Head of Data, AI & Analytics, Syngenta