Purpose

• This internship project aims to optimize EmbSW testing processes by intelligently selecting tests most likely to detect defects.
• Goal is to reduce testing time and resource consumption while maintaining high software quality.

Work To Be Done / Objectives

• Define: Analyze existing test suites and historical test execution data to identify patterns for predictive modeling.
• Design and develop: Create machine learning algorithms to predict the most relevant subset of tests to run based on recent code changes, historical failures, and risk factors.
• Enhance: Measure prediction accuracy, test coverage, and reduction in testing time to validate effectiveness.

Responsibilities / Tasks

• Collect and preprocess historical test execution logs, code-change metadata (commits, diffs), and failure records.
• Engineer features representing test, code and risk attributes; evaluate different ML models and selection strategies.
• Integrate predictive selection logic into CI/CD pipelines (e.g., Jenkins) to automatically select and trigger the most relevant tests.
• Develop evaluation metrics and dashboards to measure prediction accuracy, test coverage, time savings and regression detection rate.

Technical Environment & Keywords

• Relevant technologies: DevOps, AI, NLP, CI/CD, Jenkins, Git, Python, Robot Framework.
• Typical tools: Python ML libraries (scikit-learn, XGBoost, PyTorch/TF as needed), data processing (pandas), pipeline automation (Jenkins, Git hooks), test frameworks (Robot Framework).

Expected Deliverables & Evaluation

• A reproducible ML prototype that selects a subset of tests given recent code changes and historical data.
• Quantitative evaluation showing prediction accuracy, impact on test coverage, and reduction in overall testing time.
• Integration proof-of-concept with CI/CD demonstrating automatic test selection and execution.

Candidate Profile & Skills

• Strong background in machine learning and data analysis; experience with Python and ML libraries.
• Familiarity with software testing concepts, CI/CD tooling (Jenkins), and version control (Git).
• Good software engineering practices: data pipelines, reproducible experiments, and basic automation scripting.

Location & Administrative

• Site: STTunis 2026 (project location indicated as Tunis) — internship listing and application are provided by STMicroelectronics.

Link to apply and full job posting