built
- Team context layer2026
a reviewed, versioned memory every analyst's AI reads before it works
Claude Code · git · Python · GitHub Actions
- Workforce quality assurance pipeline2025
AI investigators that review an entire provider workforce, every case, in full
LLMs · dbt · Snowflake · Python
- Billing integrity pipeline2026
LLM classification plus per-person baselines on every timesheet, every weekday
LLMs · statistics · dbt · Python
- Model bench2026
cross-vendor evals with frozen labels that rerun from cache on every release
OpenAI · Anthropic · Python
- Billing knowledge eval2026
do frontier models actually know US healthcare billing rules?
evals · CMS rules · Python