The problem
Document-heavy workflows often depend on people moving files, renaming them, splitting or merging PDFs, applying security, and delivering the result through different channels. The manual steps are slow, inconsistent and hard to prove later.
What I built
Steller PDF evolved into a workflow appliance that can intake jobs, validate PDFs, process them through defined operations, route outputs, track job state and recover safely after interruption.
Engineering details
Capabilities include watched-folder and queue-based processing, duplicate prevention, PDF validation, merge/split/rotate operations, image conversion, barcode-driven workflows, AES-256 PDF operations, SFTP/FTPS/webhook/S3-compatible delivery, audit evidence, encrypted recovery packages, signed updates and rollback.
Why it matters
The project demonstrates the same pattern I use in client automation: move critical work out of tribal knowledge, make state explicit, make failures visible, and build recovery into the workflow instead of bolting it on later.
Shipped working software with a production-oriented operating model, not a one-off script.