Computer Vision Workshops

Seeing clearly through code.

Teams-central brings practical computer vision programming to learners across the state - working through real image data, real model failures, and real fixes, session by session. Distance is not a barrier here.

Computer vision workshop session with participants reviewing model output
Participant working through an image classification exercise
Code review during a computer vision group assignment

Who finds this useful - and who probably won't

Good fit You write Python regularly. You've touched NumPy or PIL. You want to go further with images and models.
Working developers

You build software but haven't worked with visual data. These workshops add a concrete, testable skill to what you already do.

Python background required No ML degree needed Remote-first setup
Graduate students

Research projects often need vision pipelines. Working through assignments here gives you hands-on fluency faster than reading papers alone.

Assignment-based pace Peer feedback included Flexible scheduling
Not a match yet

If you haven't written a Python function before, the pace here will feel steep. A short Python foundations course first would help significantly.

No absolute beginners Prereq: basic scripting Prep resources linked

The people around you matter as much as the material

Workshops run inside dedicated Microsoft Teams channels - ms teams spaces where your cohort shares code, asks questions, and reviews each other's outputs between live sessions. The msteams environment keeps everything in one place: recordings, assignments, feedback threads.

Participants come from across Michigan and beyond, which means your peers are dealing with the same connectivity constraints and time-zone considerations you are. That shared context shapes how the community communicates.

18+ states represented in current cohorts
4 live sessions per workshop track
8 max participants per cohort
Participants collaborating in a microsoft teams workshop channel
Computer vision model output reviewed during a group session

When this format actually works for someone

1
You can commit 5–7 hours weekly

Assignments between sessions take real time. Skimming them produces thin results - the exercises are designed to surface confusion, not reward speed.

2
You have a working Python environment

Sessions move quickly. Having conda, pip, and a GPU-accessible setup (local or cloud) ready before day one removes friction that derails early progress.

3
You're comfortable asking questions in writing

Most support happens asynchronously through the ms teams channel. Participants who write out their confusion clearly get faster, more useful responses.

4
You don't need guaranteed outcomes

Progress depends on what you bring to each session. The structure helps, but it doesn't substitute for the work itself.

Workshop participant setting up a computer vision pipeline on their workstation

A moment that comes up more than once

Participant Dagny Vreeland who completed the object detection workshop track

Dagny Vreeland joined the object detection track while working a full-time data engineering role. She had used Python for ETL pipelines for years but had never touched image data. Three sessions in, her model was detecting defects in manufacturing images she'd pulled from her own job. It wasn't perfect - precision was low on edge cases - but it was running, and she understood why it was failing.

That gap between running code, understanding failures, and fixing them deliberately is where most of the real learning in these workshops happens.

Dagny Vreeland Data engineer, object detection track participant

What's available when you're stuck

01 Async thread in microsoft teams

Each cohort has a dedicated msteams channel. Post your error, your code snippet, your confusion - responses come from instructors and peers, usually within a day.

02 Office hours (twice per track)

Two optional live sessions per track where you can bring a specific problem. Not mandatory - useful when async back-and-forth isn't resolving something.

03 Assignment walkthroughs

After submission deadlines, recorded walkthroughs show one working approach to each assignment. Comparing your solution to another working one often clarifies more than written feedback alone.

Instructor walkthrough of a computer vision assignment solution

Where you are versus where this takes you

Right now, you probably know how to write Python that processes structured data. After working through a full track here, you'll have built pipelines that process images - classifiers, detectors, or segmentation models - and debugged them on real datasets.

That's a concrete, demonstrable shift. Not a certificate that says you can, but code you wrote that does. The gap between those two things is what these workshops are designed to close, one assignment at a time.

Founded in 2016, Teams-central has been connecting geographically dispersed learners to hands-on technical instruction - before remote learning became the default, and long after it stopped being novel.

See the full program
Instructor reviewing computer vision model results with a workshop participant