INESC TEC is accepting applications to award 1 Research Grant for Degree holders - BringTrust - HASLAB (AE2026-0348)

Inesc Tec; Portugal; On-site; EUR 11,000 - 13,000; Full time

Benefits offered by this job

Health insurance

Job summary

INESC TEC invites grant applications for a Research Grant (BI) in Computer Science, focused on creating realistic simulated environments for CI/CD workflows via GitHub Actions. The project leverages large language models to generate synthetic repositories and data, preserving structure and dependencies while protecting real code and credentials.

The grant lasts 6 months starting 2027-01-01, with a monthly maintenance stipend of 1090.98 EUR and potential supplements up to 50% of the stipend,

Qualifications

• Grant duration: 6 months, starting 2027-01-01.
• Maintenance stipend: 1090.98 EUR per month, paid via bank transfer.

Responsibilities

• Design and implement mechanisms for automatically generating realistic simulated execution environments for CI/CD workflows.
• Explore use of LLMs to generate synthetic repositories and data while preserving structure and dependencies.
• Produce synthetic yet realistic repositories with plausible artificial secrets for safe analysis.

Job description

Organisation/Company INESC TEC Research Field Computer science » Informatics Researcher Profile First Stage Researcher (R1) Application Deadline 25 Nov 2026 - 23:59 (UTC) Country Portugal Type of Contract Temporary Job Status Full-time Hours Per Week 36 Offer Starting Date 1 Jan 2027 Is the job funded through the EU Research Framework Programme? Not funded by a EU programme Reference Number AE2026-0348 Is the Job related to staff position within a Research Infrastructure? No

CALL FOR GRANT APPLICATIONS (AE2026-0348)

INESC TEC is now accepting grant applications to award 1 Research Grant (BI) within the scope of the project BringTrust, with reference 18493 (COMPETE2030-FEDER-01489500) Co-funded by ERDF - European Regional Development Fund through the Innovation and Digital Transition Thematic Programme (COMPETE 2030) within the scope of Portugal 2030.

1. GRANT DESCRIPTION

Type of grant: Research Grant (BI)

General scientific area: COMPUTER SCIENCE

Scientific subarea: Informatics

Grant duration: 6 months, starting on 2027-01-01 , with the possibility of being renewed until the end of the project.

Maintenance stipend: 1090.98, according to the table of monthly maintenance stipend for FCT grants ( https://www.fct.pt/wp-content/uploads/2024/02/Tabela-de-Valores-SMM_atu… ), paid via bank transfer. Grant holders may be awarded potential supplements, according to a quarterly evaluation process (Articles 19, 21 and 22 of the Regulations for Grants of INESC TEC and Annex II), up to a maximum limit of 50 of the monthly maintenance stipend.

The grant holder will benefit from health insurance, supported by INESC TEC.

2. OBJECTIVES:

The main objective of the work to be developed during the grant is the design and implementation of mechanisms for automatically generating realistic simulated execution environments for the dynamic analysis of CI/CD workflows based on GitHub Actions. The work aims to explore the use of large language models (LLMs) to generate synthetic repositories and data that preserve the structure, dependencies and characteristics of the original repository, as well as artificial secrets, so that the workflows under analysis behave as they would in the real environment without ever accessing real code, data or credentials.

Such a solution will be essential to safely execute and observe potentially malicious third-party code, analyzing malicious behaviors such as secret exfiltration or code tampering without exposing organizations’ intellectual property. The results will help strengthen software supply chain security, an increasingly frequent target of cyberattacks.

3. BRIEF PRESENTATION OF THE WORK PROGRAMME AND TRAINING:

The work plan for this grant includes the following main activities:

• 1. Repository characterization and state of the art: study of existing techniques for synthetic code and data generation, particularly those based on LLMs, and definition of a profile of the original repository (languages, directory structure, dependencies, build and test configurations, and secrets used by the workflows) to guide the generation of the simulated environment without revealing its contents.
• 2. Automated generation of simulated environments with LLMs: design and implementation of an LLM-based generator that, from the profile of the original repository, produces synthetic yet realistic repositories, with the same structure, dependencies and build configurations, and plausible artificial secrets. Automation of the provisioning of the analysis environment, using local virtual machines or containers as epistemic GitHub Actions runners, configured through infrastructure-as-code tools (e.g., Vagrant, Ansible), so that each environment is created from scratch with the generated repository and secrets and destroyed after the analysis.
• 3. Integration and experimental evaluation: integration of the generator into the isolated and epistem