Build a Simple Neural Network to Read Handwritten Postal Codes
Overview
What this challenge is about.
You receive a labeled dataset of about 60,000 handwritten digit images (28x28 grayscale) drawn from Indian postal forms. Build two models from scratch in PyTorch: (1) a 2-layer multi-layer perceptron baseline, (2) a small convolutional neural network (CNN). Train both with proper validation splits and regularization (dropout or weight decay). Report top-1 accuracy on the holdout set, confusion analysis, and per-digit recall (since a 1-in-10 digit error in a 6-digit code is a full mis-routed parcel). Write a 2-page brief comparing your CNN result to a quoted vendor accuracy of 99.2%.
The Brief
What you'll do, and what you'll demonstrate.
Train and evaluate a small neural network for handwritten digit recognition that the operations team can compare to a vendor OCR module.
Earning criteria — what you'll demonstrate
- Implement a feed-forward and a convolutional neural network from scratch
- Use validation splits, early stopping, and regularization correctly
- Reason about per-class errors instead of relying only on top-1 accuracy
- Translate a model evaluation into a business comparison
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
Implementing and comparing baseline neural networks in PyTorch with proper validation discipline is the bread-and-butter of an entry-level MLE role.
This challenge sharpens
- neural-networks
- pytorch
- regularization
Computer Vision Engineer
A CNN on a real OCR problem is the smallest possible computer-vision portfolio piece and a credible interview talking point for junior CV roles.
This challenge sharpens
- convolutional-neural-networks
- pytorch
- model-evaluation
AI Engineer
Comparing in-house vs. vendor with both accuracy and cost framing is exactly the kind of build-vs-buy memo AI engineers write in their first year.
This challenge sharpens
- python
- model-evaluation
- neural-networks