PhD Measuring and Scaling Agentic Software Engineering (f/m/x)
Core
Design benchmarks to evaluate coding agents on real automotive codebases, investigate failure modes in code preservation, and develop verification/testing approaches for agent-generated code.
Role type
PhD researcher in agentic software engineering
Builds
Research and deployment frameworks for measuring and scaling agentic software engineering
Domain
Automotive industry + Agentic AI/LLMs
Deliverable
research
Required skills
Machine learning, Large language models, Agentic systems, Python programming, Agent evaluation and benchmarking
Preferred skills
C/C++, Java, Go programming, Experience with SWE-bench/Tau-bench frameworks, First publications at scientific venues
Technologies
Python, C/C++, Java, Go, LLMs, SWE-bench, Terminal-Bench, tau-bench
Responsibilities
Design benchmarks for coding agents, investigate code loss between exploration and generation, develop verification and testing approaches, benchmark agent configurations and LLMs across heterogeneous stacks, translate findings into research frameworks
Seniority
PhD candidate
