Software Engineering InternAt — Google

Join an agile software team focused on improving code quality and reliability generated by Large Language Models (LLMs).

Tasks:

The project description is as follows:

Large language models have shown great promise in terms of project-wide refactors and code understanding. The Linux Kernel is one of the most daunting and formidable codebases for an LLM as it is riddled with domain-specific technologies and legacy code (stuff the LLM likely wasn't trained on).

This position requires researching the practicality of using LLMs to generate real and useful security patches for the Kernel. In doing so, we hope to show the greater kernel community the usefulness of agentic systems.

Training:

- All the typical HR trainings
- IP safety training
- Insider trading training
- Git training
- Python, C++ trainings
- AI systems trainings (LLM systems)

Learning Outcome:

- work effectively with a small team of software engineers
- reason about complex tasks
- thrive in ambiguity
- understand the importance of secure AI systems in the modern computing age
- deploy production-ready code to the world's largest distributed computer system (Google infra).