Data pipeline
Course catalogue, scraped and structured
An n8n workflow pulls the catalogue, sorts it, and creates every record in one run.
01Selected work
Systems in production.
Knowledge system
The library it produces
Filterable by provider, rating and level. Rerun the pipeline and it stays current.
Cross-service automation
Folders and review links, created automatically
A webhook checks what exists, builds the missing piece in Drive or Frame.io, and writes the link back.
Pipeline and tracking
Recruitment built on relations
Roles, candidates and stages as connected databases. Position views update themselves.
Connected workspace
An operating environment for a company
Eight departments over one shared core, with access handled properly.
Reporting automation
One trigger, four reporting branches
Figures aggregated per branch and written straight into the reporting system.
Reporting
Project status you can actually see
Status charts driven off the live project data, not a manual summary.
Integration
CRM to spreadsheet, running daily
New contacts land where the team needs them. Running in production.
Eight of a larger set
How the course pipeline works
The client needed a curated, searchable body of AI and data training material. Assembling it by hand would have taken weeks and been out of date immediately, so we built the collection process instead of the collection.
The workflow requests the source catalogue, passes each result through a sorting step, and then splits into several branches that each create records in the destination library. One run builds the whole thing. When the source changes, you run it again.
How the folder automation works
The marketing team was creating a storage folder, setting up a video review space, and pasting both links back into their tracker by hand, for every single piece of work.
Now a webhook fires when the record is created. A router checks what's already there and takes one of two paths: build the Drive folder, or build the Frame.io space and generate a share link. Either way the link is written straight back to the record that started it. Nobody copies a URL any more.
Why the recruitment tracker is built on relations
Recruitment runs on the relationship between three things: the roles you're filling, the people you're considering, and the stage each pairing has reached. In a spreadsheet, that relationship is maintained by hand and drifts within weeks.
Here, postings, candidates and employees are separate connected databases. Vacant and occupied positions aren't lists anyone updates — they resolve from the underlying status. Move a candidate a stage and every view that depends on it updates too. The records shown are placeholders.
