Cayler Miley

Experience

At Redwood Materials, I work on computer vision and ML systems for battery recycling: classification pipelines for sorting materials, deployed on edge hardware and orchestrated through Kubernetes. The work spans model development, data infrastructure, and production deployment.

Previously at Clear Capital, I built multi-modal valuation models combining property images and structured tabular data to power automated property assessments. I also worked on accuracy improvements and explainability metrics for the automated valuation model (AVM).

ML Interests

I love the sequence of model research, development, and then production. I regularly use off-the-shelf vision models (DINOv2, SAM, YOLO, SigLIP), VLMs (QWEN), and OCR (HunyuanOCR), often with custom variations to better fit data and business outcomes. I also enjoy reading new research and working with traditional modeling approaches (XGBoost, LightGBM, ARIMA). I’ve previously spent time on projects related to biologically inspired computing (via spiking neural networks) and speech recognition.

Outside of Work

I play guitar: metal, jazz, and whatever else is flavor of the month. I golf, play video games, and enjoy a good whiskey. In the summer you can find me camping, boating, and occasionally riding motorcycles.