Products we've built, pipelines we've engineered, and tools we've shipped — for clients and ourselves.
Claude Code lets teams deploy agents, skills, and commands — but visibility into what's actually working stops there. Claude Code Telemetry closes that gap. It tracks which agents and skills get invoked, what's driving token spend per workflow, and how adoption breaks down across your team. Engineering leaders finally have the data to optimize their AI tooling investment — not just deploy it. Now in early access.
Read case study →The Transparency in Coverage mandate made healthcare pricing data public. Actually working with it is another story. We built a hybrid Python pipeline that streams and processes hundreds of gigabytes of United Healthcare machine-readable files without loading them into memory — using browser automation for file discovery, DuckDB and Athena for querying, and a Next.js dashboard for analysis. A research team got access to data they'd never been able to query before, delivered in days at a fraction of what traditional ETL infrastructure would cost.
Read case study →spaget.ai is where engineering teams build their AI development toolkit. Design agents, skills, commands, and hooks in one place, then deploy them across Claude Code or any other AI coding environment. It's infrastructure for teams serious about AI-assisted development — giving you the flexibility to compose and share capabilities without starting from scratch every time the tooling landscape shifts.
Read case study →Additional work from our portfolio
A unified AWS dashboard that consolidates resource inventory, service metrics, and billing into one clean interface — instead of five separate consoles.
An agent-based simulation for teaching and exploring healthcare insurance economics — how individual agents face health events and interact with different payer models over time.