Segment Cells from Microscopy Images for a Pharma-AI Discovery Lab
Overview
What this challenge is about.
You receive 3,500 microscopy images with pixel-level cell masks plus a 200-image hold-out set re-annotated by two biologists for inter-annotator agreement. Train a U-Net or SegFormer model. Evaluate IoU (Intersection over Union) and Dice score against the hold-out, and compare to the legacy thresholding pipeline. Report failure modes (touching cells, debris, out-of-focus regions). Deliverable is the trained model, the comparison report, and a 2-page handoff brief.
The Brief
What you'll do, and what you'll demonstrate.
Beat the legacy thresholding pipeline on IoU and cut the weekly manual-fixup time, while staying interpretable to the biologists.
Earning criteria — what you'll demonstrate
- Train modern semantic-segmentation models on biological imagery
- Compare against a legacy production pipeline fairly
- Quantify inter-annotator agreement as a context for model performance
- Communicate model behavior to a domain-scientific audience
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
Machine Learning Engineer
Replacing a legacy pipeline with a deep-learning model and shipping it to a daily-user audience is exactly junior MLE territory.
This challenge sharpens
- semantic-segmentation
- u-net
- pytorch
Computer Vision Engineer
Biological-imaging segmentation is a high-leverage CV engineer specialization at pharma-AI and lab-automation companies.
This challenge sharpens
- semantic-segmentation
- biological-imaging
- evaluation
Applied AI Scientist
Quantifying inter-annotator agreement and writing biologist-facing briefs is the applied-AI-scientist skillset that bridges research and lab teams.
This challenge sharpens
- annotation-agreement
- evaluation
- biological-imaging