Build a Real-Time Cascade Detection Pipeline for a Trading Platform
Overview
What this challenge is about.
Receive 6 months of anonymized trade events (timestamp, ticker, side, volume, account_segment) plus a static follower graph among 50k 'top trader' accounts. Build a streaming pipeline (Kafka or Redpanda + Python consumers) that computes: rolling order-flow imbalance per ticker, follower-influence-weighted volume (volume of trades by accounts followed by N influencers), and a graph-cascade-likelihood score over 5-minute windows. Back-test on 12 known cascade events and 24 quiet windows; report precision/recall at 3 alert thresholds. Deliver: 16-page architecture doc, working prototype on historical replay, and a 6-page risk-desk runbook.
The Brief
What you'll do, and what you'll demonstrate.
Build a real-time network-based cascade detection pipeline and back-test it against 12 known events with publishable precision/recall.
Earning criteria — what you'll demonstrate
- Build a streaming pipeline that joins event data with a static graph
- Combine order-flow and network features into a single cascade score
- Back-test detection on a small labeled-event sample honestly
- Translate a detection score into operational risk-desk action
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.
Career mappings coming soon.