The situation
The robotics lab teaches around 40 graduate students per cohort; a clean teaching reference saves about 200 staff-hours of TA debugging support per semester.
Your task
Ship a reproducible end-to-end lane-following teaching reference with 3 ablations and a student-ready lab guide.
The brief
What you do
- Use the public Donkeycar Tub dataset (or collect about 30 minutes of driving on the simulator).
- Train a CNN-policy baseline (the Donkeycar default architecture is fine) that predicts steering + throttle from camera frames.
- Run 3 ablations: image augmentation on/off, RGB vs. grayscale input, and a smaller backbone.
- Evaluate via simulator lap-time and lane-departure count.
You deliver
A teaching notebook + 5-page lab guide for next semester's students.
What you will practice
- Train an end-to-end imitation-learning policy from camera data
- Design and run informative ablations
- Evaluate driving models in simulation with clear metrics
- Document a teaching reference that students can rerun
A suggested way in
- Donkeycar gym is the standard simulator for the platform
- Hold out a separate driving session as the evaluation set; never tune on it
- Report lap-time AND lane-departure count — one number lies
Skills you will practice
End to end learning, Imitation learning, Pytorch, Data augmentation, Ablation study, Documentation
Where it leads
AI Engineer, Machine Learning Engineer, AI Product Designer
How it is assessed
- Reproducibility30%
Notebook + lockfile let a student reproduce results on a clean install.
- Ablation informativeness25%
Ablations clearly isolate the effect of each variable.
- Driving performance25%
Best ablation completes 3 sim laps with zero lane departures.
- Lab-guide quality20%
A first-week graduate student can complete the lab from the guide alone.
Resources
- PyTorchTool
Train the small CNN policy.
- Donkeycar gymTool
Reference simulator the course already uses.
- JupyterTool
Notebook format students can step through.
- MatplotlibTool
Plot training curves + ablation results.
- NotionTool
Author the student lab guide.
- GitTool
Version the teaching reference for next semester's drop-in.
Finish it. Earn this.
A verified certificate with your name. Anyone can check it.
Time to practice.
Solve this challengeMore like this.
- Software EngineerBuild a Multi-Level Feedback Queue Scheduler in xv6
- Software EngineerReverse-Engineer a Stripped Binary to Restore Linking Symbols
- Software EngineerImplement a Distance-Vector Routing Algorithm
- NLP EngineerLLM-Powered FAQ Chatbot for 40-Person SaaS Scale-up
-
Build a Multi-Level Feedback Queue Scheduler in xv6 Software Engineer Fork the xv6-riscv repository and implement MLFQ in proc.c with 3 priority queues, per-queue time slices (10ms / 20ms / 40ms), priority demotion on sl operating systemsschedulingc programming Open challenge → -
Reverse-Engineer a Stripped Binary to Restore Linking Symbols Software Engineer Receive the stripped shared object (Linux ELF, x86-64, ~280KB), a crash dump with 12 frames of indecipherable addresses, and 4 years of partial git hi reverse engineeringlinking and loadingmachine code Open challenge → -
Implement a Distance-Vector Routing Algorithm Software Engineer Implement distance-vector routing in Python: each simulated router maintains a distance vector, exchanges vectors with neighbors every 1 second of sim routingdistance vectornetwork simulation Open challenge → -
LLM-Powered FAQ Chatbot for 40-Person SaaS Scale-up NLP Engineer You have access to TaskFlow's internal documentation, help articles, and a sample of 500 support tickets. Your task is to build a retrieval-augmented llmraginformation retrieval Open challenge →