Overview
VisionHacks is a one-day, in-person computer vision hackathon on Saturday, December 5, 2026 at Sheridan College in Oakville, co-hosted with Inference Labs.
Industry partners share real use cases, like quality inspection, safety monitoring and inventory tracking. Teams of up to 4 build working computer vision prototypes on Inference Labs' Sertn platform, competing for a $3,000 prize pool.
Get started
- Apply at visionhacks.com/apply. Only accepted hackers can participate.
- Form a team of 1–4. Come solo and we'll help you match on the day.
- Before the event, watch the Sertn technical intro video and set up your Sertn access.
Schedule
- 9:00 AM: Check-in & breakfast
- 9:30 AM: Opening & challenge reveal
- 10:00 AM: Hacking begins
- 12:30 PM: Lunch
- 3:30 PM: Submissions due
- 3:45 PM: Demos
- 5:00 PM: Judging
- 5:30 PM: Awards & closing
Requirements
What to Build
A working computer vision prototype that tackles one of the partner challenges revealed at the opening.
What to Submit
- Project name, a short description, and which partner challenge you chose
- A link to your code repository
- Optional: a 1–3 minute demo video and any deployment links
⚠️ Submissions close at 3:30 PM. Leave at least 30 minutes for your Devpost submission; live demos start at 3:45 PM.
Prizes
$3,000 Prize Pool
Judges
Ron Chan
Co-Founder @ Inference Labs
Elena Pashkova
CMO @ Inference Labs
Judging Criteria
-
Challenge Fit
Pass or Fail — Does the project address one of the partner challenges? -
Business Value
30% — Does it solve the partner's problem in a way that saves time, cost or risk? Would the company want to use it? -
Technical Execution
30% — Does the computer vision pipeline work on realistic inputs? How accurate and robust is it? -
Deployability
20% — Could it realistically be piloted? Is it practical in speed, cost and integration, and does the team understand its limits? -
Presentation
20% — Was the demo clear and convincing? Did the team explain their approach, results and trade-offs?
Questions? Email the hackathon manager


