Overview

The AI/ML Project Evaluator assesses team submissions in Artificial Intelligence and Machine Learning competitions against the project Software Requirements Specification (SRS) and a 100-mark evaluation rubric. The evaluator is responsible for reviewing 25 projects and verifying that each project works as expected, rather than simply checking whether the required features are present. This includes running the submitted applications, validating datasets and preprocessing steps, checking model performance against the actual code and results, comparing Python ML and Google Teachable Machine models, and recording fair, evidence-based marks and feedback.

Key Responsibilities

Each evaluation follows a structured process, from receiving the project submission to recording the final marks and feedback. The evaluator must review each project against the requirements defined in the SRS and assess the implementation, model quality, testing, documentation, and overall project readiness.

The evaluator must verify that the submitted functionality works correctly, model results are supported by actual evidence, and the evaluation criteria are applied consistently across all teams.

Evaluation Areas

The evaluator scores 11 criteria with a combined maximum of 100 marks.

Functionality & SRS Compliance

Verifies that all required features are implemented and working as specified in the SRS, including correct inputs, outputs, workflows, and error handling.

Dataset Quality, Pre-processing & Feature Engineering

Checks dataset relevance and quality, missing values, duplicate records, data cleaning, preprocessing, feature selection or extraction, and appropriate preparation of training and testing data.

Python Machine Learning Model Quality

Reviews the Python ML implementation, appropriate algorithm selection, training and testing procedures, prevention of data leakage, suitable evaluation metrics, model performance, and reproducibility of results.

Google Teachable Machine Model Quality

Checks whether the Teachable Machine model is trained correctly, uses appropriate examples and classes, performs well on unseen inputs, and is integrated into the project as required.

Dual-Model Comparison, Rule Validation & Hidden Testing

Verifies the comparison between the Python ML and Google Teachable Machine models, checks whether the comparison uses actual results, validates required business rules or decision rules, and tests the application using unseen inputs and hidden test cases.

Source Code Quality & Technical Implementation

Reviews code structure, readability, modularity, maintainability, reproducibility, dependency management, error handling, security, and removal of unnecessary or duplicate code.

UI, Dashboard & User Experience

Checks usability, navigation, input and output presentation, model predictions, error messages, dashboard functionality, and whether the interface meets the project requirements.

Testing, Performance & Scalability

Reviews test coverage, correctness of results, response times, handling of different inputs, and performance under increased data volumes or usage where applicable.

Documentation & Published Blog

Verifies the accuracy and completeness of project documentation, system architecture or workflow diagrams, installation and execution instructions, model details, published blog, and demonstration video.

Plagiarism & AI-Usage Compliance

Checks originality, appropriate disclosure of AI-assisted work, completeness of the AI usage document, and evidence of individual team contributions through version control where required.

On-Time Submission

Verifies whether the required source code, project files, documentation, final commits, and other deliverables were submitted before the deadline.

Required Skills

The evaluator must have practical AI/ML knowledge and the ability to assess technical implementations, validate model results, and provide fair, evidence-based feedback.version control systems, software testing, and basic application security.