| title | GitHub Copilot for data scientists |
|---|---|
| description | A five-slide briefing on Copilot in VS Code, HVE RPI and model selection. |
A short briefing for Desjardins teams using GitHub Copilot for data science in VS Code. The colors are Desjardins-inspired, not an official brand template.
- Lire la présentation Web
- Télécharger le PowerPoint de cinq diapositives
- Télécharger le PDF en français
- Read the English web presentation
- Download the five-slide English PowerPoint
- Download the English PDF
The five sections cover the data-science workflow, completion/Ask/agentic interactions, the HVE Core RPI lifecycle, model choice, and a synthetic credit-risk example. Each PowerPoint includes presenter notes and source links in its own language. French and English remain separate; the site opens in French and offers a language switch.
Open the French page or the English page directly, or serve the repository with a static HTTP server. No build step, external libraries or network-loaded fonts are required. The page supports mobile layouts and works as a readable document without JavaScript.
Select Présenter (French) or Present slides (English) for one section at a time. Use the previous/next buttons or the left/right arrow keys to navigate. Press Escape or select Lire la page / Read page to return to the full page. Download the PDF for a fixed five-page print version.
This repository already uses GitHub Actions as its Pages source.
The Pages workflow publishes only the web page and
presentation downloads after a push to main, or when run manually.
Commit and push these files to publish the new content; local changes are not live.
For another repository, set Settings > Pages > Build and deployment > Source
to GitHub Actions first.
Auto routes among available models subject to your plan and administrator policies. Hydrafusion is setup-dependent, not a universal Copilot option. Large, medium and small refer to relative capability recommendations, not product labels, parameter counts or reasoning-effort settings.
Use synthetic or approved de-identified data for the example. Do not include client PII, credentials or confidential datasets in prompts. Copilot output still requires technical review and the customer's model-risk controls.