AI farming: 100x the yield with a data team of 1
Bowery Farming is building and operating the world's most technologically advanced indoor vertical farms. Our farms combine robotics, hyper-granular sensing & control, computer vision, and machine learning to reduce water use by 95% and increase yield by 100x compared to traditional agriculture. For the last year we have pursued the data-centric approach to agriculture with a data-eng / analytics / ML team of ... 1. Thanks to the amazing progress in cloud infrastructure and open source warehouse tools, we've been able to prove the value of these systems before hiring the full team to dig deeper. Our software architecture is built around streaming event data and our stack includes AWS SQS, Kinesis Firehose, and Redshift with data modeled in dbt and shared internally with Mode. Our production data products use python, R, H2O, TensorFlow, and Google's OR-Tools. I'll describe how we use these tools to build data infrastructure which is lightweight yet scalable, adaptable yet reliable.
