ANALYTICS & AI
AI Running Coach
Running analytics, a visual dashboard and opt-in AI coaching, with reproducible calculations kept separate from generated advice.
- FastAPI
- PostgreSQL
- JavaScript
- OpenAI
Personal projects. Practical engineering.
Hi, I'm Gianpaolo. I build applications around real-world problems, connecting interfaces, APIs and data with a hands-on, self-hosted approach.
01 / SELECTED BUILDS
A few examples of the applications I develop. App links live in the launcher below and may require sign-in.
ANALYTICS & AI
Running analytics, a visual dashboard and opt-in AI coaching, with reproducible calculations kept separate from generated advice.
FULL-STACK APPLICATION
Vehicle records, maintenance, fuel spending and mileage brought together through a relational API and an interactive web interface.
AI & DOCUMENT AUTOMATION
Job-specific CV generation from a verified source profile, with optional AI rewording and checks for unsupported numbers and names.
02 / HOW IT FITS TOGETHER
Two connected flows: getting code into a running container, then serving an application and its data.
Repository-defined workflow, not live infrastructure monitoring. On smaller screens, open full size or use the text version below.
Open full-size diagram (opens in a new tab)From a source change to a deployment check
Application source and a checked-in Jenkinsfile.
Source controlSCM polling picks up changes and builds the Docker image.
Build automationThe job replaces the app container with the new build.
DeploymentThe pipeline requests an app endpoint to check the deployment.
VerificationSelected jobs also register the app. For example, CV OnDemand updates the live service registry after its health check. The launcher reads that registry.
From a browser request to app-owned storage
Open the portal or navigate to an application's address.
Web requestServes the static portal and routes requests to hosted apps.
HTTPS & routingEach hosted application runs its own interface and backend.
Independent servicesRunning Coach and My Garage use PostgreSQL. The shared-server convention gives each app its own database or schema.
MTG uses SQLite. CV OnDemand keeps YAML profiles and generated documents in app-owned file storage.
The PostgreSQL convention is a dedicated database or schema per app, not one common database for everything. Some apps use SQLite or files instead.
These Jenkins jobs include build, deploy and endpoint checks. Tests exist in individual projects; a universal pipeline test gate is not shown here.
The launcher reads the service registry. Portal deployments preserve its live copy; app registration is a separate operation.
03 / FAMILY PORTAL
Browse the configured apps and infrastructure. Links open separately; access is handled by the destination, not this page.
Loading services...
Reachability is a browser-side network check, not a health or authentication guarantee. App descriptions come from the registry; HTTP-only checks are skipped when this page uses HTTPS.