Staff AI Software Engineer – Code Intelligence & Quality Validation
TikTok
- Location
- San Jose, California, United States of America
- Employment
- Full Time
- Work model
- On-Site
- Level
- Staff
- H-1B history
- 148 approvals (FY2023)
About this role
The Agentic Ops US team focuses on two major areas: Code Graph and Quality Validation. The code graph serves as the data foundation for validation, encompassing static relationships (function calls, experiment/instrumentation dependencies) and dynamic execution paths (reconstructed from traces). Quality validation covers static rule checking (mining soft and logical constraints) and dynamic issue detection, reproduction, and fixing. We aim to improve code reliability and R&D efficiency through systematic, data-driven approaches.
Responsibilities - Design, develop, and optimize Code Graph capabilities, including static relationship extraction, dynamic trace processing, execution path reconstruction, and graph data modeling. - Contribute to the architecture and technical design of core systems, with a system-level view of module boundaries, dependencies, scalability, and maintainability, and drive technical solutions through implementation. - Design and develop static validation capabilities based on the code graph, including mining and validating logical and soft constraints, while continuously improving rule accuracy and coverage. - Build dynamic validation capabilities for production issue detection, reproduction, root cause localization, and assisted fixing, and develop reusable solutions for common problem patterns. - Conduct in-depth analysis of complex production quality issues by leveraging code, traces, and other runtime signals, identify root causes, extract generalized patterns, and drive automated validation coverage. - Develop a deep understanding of code organization, module dependencies, build pipelines, and runtime behavior in large-scale client applications, and explore how Code Intelligence can improve the understandability and quality of complex codebases. - Explore and introduce techniques in code analysis, program slicing, anomaly detection, LLMs / Agents, and related areas, and apply them to real-world engineering problems. - Collaborate closely with client, backend, and infrastructure teams to dri