// proof in production
The stack changes. The pattern does not: important systems, messy operating reality, and nobody else willing to own the whole path. These are running today.
// aerospace and mro
A Washington, D.C.-area aerospace parts and MRO business had roughly 14,000 records, a 1.6 TB document store, QuickBooks, hand-built shipping templates, and no single job identity. We built one relational system from quote through delivery and connected labeling, RFID, inventory, and environmental monitoring.
15 boards · ~14,000 records · 1.6 TB of documents
Read the aerospace case study →// industrial agriculture
We built and operate a multi-database system spanning more than 500 API endpoints, 70+ data models, Windows and Android applications, warehouse kiosks, scales, scanners, and AWETA grading equipment.
500+ endpoints · 70+ models · hardware + software
See the deployment →// heavy civil
A heavy-civil contractor replaced binders, informal equipment transfers, and scattered inspection records with QR-based asset identity, equipment-specific compliance workflows, and a complete transfer history.
5 companies · 15+ sites · 200+ field personnel
Read the heavy-civil case study →Written up in full: Aerospace MRO deployment and Heavy civil equipment operations . Customers are described, never named.
// or skip us and read the code
Client deployments have to be taken on the write-up. These you can read before you talk to us.
Running in production in both our aerospace and heavy civil deployments — you can read the hardware layer itself. GPIO and I2C abstractions for embedded Linux, pluggable sensor drivers, retry and scheduling helpers, and config-driven pin mapping for BeagleBone. Real backends via libgpiod and Linux i2c-dev, plus a simulation backend so it runs on a laptop. Published as a gem, MIT.
github → // postgres · duckdb PG-Lake-MakeTurns PostgreSQL into a lakehouse: query Parquet on S3-compatible storage directly through Postgres using DuckDB's engine, backed by MinIO or SeaweedFS. Runs on a laptop for development or your own VM for production — which is the entire point.
github → // c csv_to_parquetA small CLI that converts CSV to Parquet through DuckDB, with type inference and correct handling of quoted fields, embedded newlines, and delimiters inside quotes. Unglamorous and useful.
github →// your turn
What is stuck, what it is costing the business, and what has already been tried. We will tell you plainly whether we are a fit.