Midpage

Build and scale RL environments and evaluation harnesses for long-horizon legal reasoning tasks. Own the data pipeline that converts court filings into verifiable training tasks and benchmarks for frontier AI labs.

Build and optimize data pipelines to augment case law datasets with metadata and ensure the reliability of web scrapers. Develop core services and APIs to expose data while benchmarking tasks using human and synthetic evaluations.