The situation
The MLOps company (around 90 staff, around EUR 22 million ARR) has identified dashboard usability as the second-most-cited reason for churn in last quarter's customer interviews.
Your task
Redesign the primary model-monitoring dashboard so the four core questions are answered in the order users actually ask them.
The brief
You start with
You receive realistic 30-day metric streams for three production models (a classifier, a regressor, and a ranker) including prediction volume, latency p50/p95, accuracy or RMSE on labeled samples, and feature drift scores.
What you do
- Design and prototype a single-page dashboard answering four questions in order: is the model healthy now, what changed in the last 24h, is drift building, and where to drill in.
- Build the prototype in Streamlit or a similar low-friction stack.
- Deliverable is the design rationale, the prototype, and a redlined comparison to the current dashboard.
What you will practice
- Apply visual-hierarchy principles to a dense operational dashboard
- Match chart type to data type for ML-specific metrics
- Design for the user's actual question order, not the data team's mental model
- Communicate redesign rationale defensibly to a product team
A suggested way in
- Use the inverted-pyramid layout (headline status on top, drill-down below)
- Reserve color for state change; use shape and position for steady-state information
- Test the prototype against the four-question script before declaring it done
Skills you will practice
Dashboard design, Visual hierarchy, Chart selection, ML metrics, Streamlit, Redline design
Where it leads
AI Product Designer, AI Product Manager, MLOps Engineer
How it is assessed
- Information hierarchy30%
Dashboard answers the four questions in the order users ask them, without scrolling.
- Chart appropriateness25%
Each chart type fits the underlying metric semantics.
- Prototype fidelity25%
Working prototype reflects all design decisions, not just the headline view.
- Rationale clarity20%
Design doc is convincing to a product manager and a senior MLE alike.
Resources
- FigmaTool
Sketch the layout and produce the redlined comparison.
- StreamlitTool
Ship the interactive prototype with minimal frontend overhead.
- PythonTool
Load and shape the synthetic metric streams.
- AltairTool
Strong chart-type defaults that pair well with Streamlit.
- NotionTool
Author the design rationale doc.
- GitHubTool
Version the prototype and design assets together.
Finish it. Earn this.
A verified certificate with your name. Anyone can check it.
Time to practice.
Solve this challengeMore like this.
- Software EngineerGreedy Delivery-Slot Assignment for a Munich Grocery Startup
- Software EngineerDesign a Verified UART Receiver and Loopback for an FPGA Sensor Hub
- Data ArchitectDesigning a Data Warehouse for a Renewable Energy Firm
- Database AdministratorCut a 42-Minute Trino Query to Under Five Minutes
-
Greedy Delivery-Slot Assignment for a Munich Grocery Startup Software Engineer Using the provided order history sample, rider roster, and manual dispatch baseline, design and implement a greedy slot-assignment algorithm in Python greedy algorithmsalgorithm analysispython Open challenge → -
Design a Verified UART Receiver and Loopback for an FPGA Sensor Hub Software Engineer Propose and deliver a UART receiver for Helax's sensor hub that meets the functional requirements in the requirements brief (the 'uart-requirements-br veriloguartfpga Open challenge → -
Designing a Data Warehouse for a Renewable Energy Firm Data Architect You are given sample data from three sources: energy production logs, weather data, and equipment maintenance records. Your task is to: (1) design a s data warehousingstar schemaetl Open challenge → -
Cut a 42-Minute Trino Query to Under Five Minutes Database Administrator Using the slow query, the EXPLAIN ANALYZE readout, the Iceberg table schemas, and the 30-day execution-metrics dataset provided, design an optimizatio trinosql optimizationiceberg Open challenge →