Applied AI Engineer intern

Applied AI Engineer Intern (PFE)

Topic: Evaluation and Optimization of AI-Powered Automated Grading Systems

Project applied to: GOMYTEACHER

📌 Project Overview

With the increasing use of AI in educational assessment, automated grading systems must ensure accuracy, fairness, and consistency across subjects and learner levels.

This internship aims to scientifically evaluate and improve the reliability of AI-powered grading systems used in GOMYTEACHER.

🎯 Main Objectives

• Compare AI-generated grades with human instructor evaluations

• Analyze grading performance across different subjects:

Programming (Python / JavaScript)

Data Science

Cybersecurity

Math & logic

Theoretical vs practical assessments

• Benchmark multiple LLMs:

GPT-4 / GPT-4o

Claude

Open-source models (Llama, Mixtral)

• Evaluate grading dimensions:

Score accuracy

Feedback relevance

Consistency

Bias & variance

Hallucination rate

• Propose improvement strategies:

Prompt engineering

Rubric-based grading

Hybrid grading (AI + rules + human-in-the-loop)

🧪 Methodology

• Build a gold-standard dataset using real student submissions graded by multiple instructors
• Design an AI grading pipeline (grade + feedback + confidence score)
• Run multi-LLM benchmarking experiments
• Perform quantitative evaluation (MAE, Cohen’s Kappa, correlation, variance)
• Conduct qualitative error and feedback analysis

📦 Expected Deliverables

Technical

• AI grading evaluation framework
• LLM benchmarking dashboard
• Subject-wise performance reports

Academic

• Final PFE report
• Evaluation tables & graphs
• Reproducible experiments

Business Impact

• Decision matrix: best LLM per subject
• Human validation thresholds
• Cost vs performance trade-offs

🚀 Optional Extensions

• Bias & fairness analysis
• Confidence-aware grading with human review flags
• Self-improving grader using instructor feedback

📩 How to apply

Please send your resume to amine@gomycode.com

Informations

Requirements

Skills:

AI INTERN INGÉNIEUR