You will conduct comprehensive literature reviews and State-of-the-Art (SoTA) analyses of current research on LLM context engineering, repository-scale retrieval-augmented generation (RAG), and agentic workflows for software development.
As part of setting up and maintaining agent infrastructure, you will configure and manage execution sandboxes, agent harnesses, and test environments.
To assess real-world impact, you will implement and run benchmarks on repository-level coding use cases, measuring performance, output accuracy, and developer productivity gains.
Furthermore, you will propose, prototype, and evaluate generic context pipelines for AI coding agents.
Finally, you will document scientific methodologies and findings and present your results to the research team.
Requisiti principali
Master studies in the field of Computer Science, Artificial Intelligence, Data Science, Software Engineering, or comparable
strong programming skills in Python, Git, and containerized sandbox environments such as Docker; solid understanding of LLMs, prompt engineering, context management, and AI agent architectures; hands-on experience developing, benchmarking, or experimenting with AI-powered coding assistants and repository-scale context retrieval systems
you are driven by curiosity, thrive when working independently, and bring a structured approach to solving complex challenges
for this position, your presence on-site is essential
