Build an Attention-Visualization Tool for Translation Quality Audit
Overview
What this challenge is about.
You will load a small open-source EN-FR transformer (e.g., Helsinki-NLP Opus-MT-en-fr), build a Streamlit or Gradio demo that lets the user paste English source, see the French output, and inspect (1) cross-attention heatmaps per decoder layer, (2) source-token alignment, (3) per-token output confidence. Recruit 3 professional translators (Upwork is fine) for a 30-minute usability session each. Synthesize feedback into a 3-page tool-design recommendation for the product team.
The Brief
What you'll do, and what you'll demonstrate.
Build a working attention-visualization tool for translation quality audit and validate its usefulness with professional translators.
Earning criteria — what you'll demonstrate
- Inspect and visualize attention in encoder-decoder MT models
- Translate model internals into translator-actionable signals
- Run lightweight usability studies and synthesize the feedback
- Recommend tool iterations grounded in user research
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.
AI Product Designer
Designing tools on top of model internals, validated with real professional users, is the AI-product-designer's signature deliverable at AI-forward consulting and SaaS firms.
This challenge sharpens
- tool-design
- user-research
- attention-mechanisms
AI Engineer
Shipping working demos that expose model internals is exactly the kind of glue engineering AI engineers do in their first year.
This challenge sharpens
- demo-development
- transformer
- neural-mt
NLP Engineer
Inspecting and exposing attention from real MT models is core NLP-engineer territory, especially for any team building human-in-the-loop translation tools.
This challenge sharpens
- attention-mechanisms
- transformer
- neural-mt