Opportunity brief
Overview:
Edge AI Challenge is a team-based technical competition focused on TinyML, Edge Computing and Embedded AI. Participants will receive a common Edge AI problem statement, dataset, target-platform specifications and submission requirements at the beginning of the competition.
Teams of 2–3 members will have 2 hours to develop, train, validate and optimize a lightweight machine-learning model for embedded deployment. Participants may apply suitable optimization techniques such as quantization and model compression and prepare their solution for deployment on the specified target platform.
At the submission deadline, all submissions will be frozen. The organizing team will subsequently deploy and benchmark valid submissions on a standardized microcontroller platform using identical test conditions and hidden evaluation inputs.
Evaluation will be based on:
- Model Performance / Accuracy – 45 points
- Inference Latency – 20 points
- Model Size / Resource Usage – 15 points
- Successful Edge Deployment – 10 points
- Technical Approach & Documentation – 10 points
The competition emphasizes practical Edge AI development and the engineering trade-offs involved in achieving strong model performance while meeting embedded hardware constraints.
Competition Duration: 2 hours Team Size: 2–3 members