AI-powered crop inspection and plant health assessment platform
Bloomfield applies computer vision and deep learning to automate crop inspection, replacing manual visual assessment with algorithmic consistency. The stack reveals a data-heavy operation: PyTorch + TensorFlow for model training, AWS infrastructure (SageMaker, Lambda, Redshift, Athena, Kinesis) for data pipelines, and active work on image processing models and production integration. The engineering-to-data ratio (6:2) and senior-heavy seniority mix suggest a mature ML-first organization focused on scaling inference and turning plant imagery into reliable data products.
Bloomfield automates crop health assessment using AI and deep learning, targeting the inspection bottleneck in agriculture. Manual crop inspection—the current industry standard—relies on trained inspectors to visually identify plant characteristics and anomalies with consistency challenges around time, labor availability, and accuracy. The company's platform captures crop imagery and applies deep learning models (built on PyTorch and TensorFlow) to detect and assess the same plant characteristics human inspectors evaluate. Active project work spans model retraining, image processing pipelines, citrus weight estimation, and cloud-based data automation. Founded in 2019 and based in Pittsburgh, the company is a public entity with 11–50 employees.
Bloomfield builds models with PyTorch, PyTorch Lightning, and TensorFlow, supported by MLflow for tracking and SageMaker for production deployment.
AWS is the primary cloud provider. The stack includes SageMaker (model hosting), Lambda (serverless compute), Redshift and Athena (analytics), Kinesis (streaming), and Glue (ETL).
Specialties include grapes and blueberries. Active work includes a citrus weight estimation AI product, indicating expansion into citrus crops.
Other companies in the same industry, closest in size