Overview
What this challenge is about.
You receive a CSV of around 80,000 accounts (existing customers + prospects) with attributes (industry, size, tech stack, geography) plus 18 months of marketing-touch and conversion data. Construct a graph with company nodes and edges for shared technologies, shared decision-makers (anonymized hashes), and industry overlap. Train a node-embedding model (Node2Vec or a GraphSAGE link-prediction head) and recommend top-5 similar accounts for each existing customer. Backtest against held-out conversions: how often does the truly-converted prospect appear in your top-5? Write a 2-page memo for the sales operations director.
The Brief
What you'll do, and what you'll demonstrate.
Build an account-similarity link-prediction model that meaningfully improves account-expansion conversion rate over hand-curation.
Earning criteria — what you'll demonstrate
- Construct an account graph from tabular CRM-like data
- Train and evaluate node embeddings for link prediction
- Backtest a recommender against historical conversions
- Communicate model uplift to a sales operations 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.
Data Scientist
Shipping a link-prediction recommender backed by a time-aware backtest and a sales-ops memo is exactly the day-one work of a data scientist in B2B SaaS.
This challenge sharpens
- link-prediction
- node-embeddings
- evaluation
Machine Learning Engineer
Building a reproducible graph + embedding pipeline that the sales team can re-score against is core MLE work for go-to-market data platforms.
This challenge sharpens
- pytorch-geometric
- graph-construction
- node-embeddings
Applied AI Scientist
Translating model metrics into a sales-ops recommendation pilot is the applied-AI-scientist craft of putting ML into operations.
This challenge sharpens
- link-prediction
- evaluation
- node2vec