Teams at Midpage

Build and scale RL environments, evaluation harnesses, and data pipelines that convert millions of court filings into contamination-free benchmarks and training tasks. Integrate with partner harnesses and model APIs, collaborate with attorneys to translate legal workflows into scorable tasks, and run large-scale, sandboxed evaluations.

Build and maintain pipelines and web scrapers to create a large case-law dataset, implement metadata extraction (citations, overruling/appeals), ensure scraper reliability, optimize core utilities, expose data via APIs/websockets, and benchmark/evaluate tasks.

Build and maintain pipelines and services to create a comprehensive case law dataset: augment documents with legal metadata, ensure reliable web scrapers, optimize citation and court-resolution utilities, expose data via APIs/websockets, and benchmark core tasks.