Research Engineer Intern, Agentic Systems & AI Infrastructure (TikTok-Generalized Arch) - 2027 Summer
Core
Research and deploy reliable long-horizon AI agents that reason over large codebases, interact with complex tools, and complete consequential engineering work at TikTok scale.
Role type
Research Engineer Intern (Agentic Systems & AI Infrastructure)
Builds
AI-native engineering infrastructure for agent orchestration, evaluation, and reliable execution across distributed production systems.
Domain
Artificial Intelligence / Software Engineering / Distributed Systems
Deliverable
production ML models
Required skills
Agent planning and reasoning, model post-training, reinforcement learning, memory and context systems, tool learning, multi-agent systems, agent evaluation, Python/C++/Go/Java programming, experimental design and analysis
Preferred skills
Research publications (NeurIPS, ICML, ICLR, etc.), experience building/deploying LLM agents or agent platforms, hands-on with execution-feedback loops and distributed agent runtimes, experience with CI/CD and production operations
Technologies
Python, C++, Go, Java, LLMs, Reinforcement Learning, Distributed Systems
Responsibilities
Conduct frontier research on reliable long-horizon agents for multi-step software engineering tasks; Advance agent capabilities in hierarchical planning, reasoning, and tool use; Co-design agents and models through post-training and reinforcement learning; Build scalable agent infrastructure for orchestration and observability; Apply new methods in workflows like cross-repository changes and incident diagnosis; Translate research into production systems and define rigorous evaluation methodologies; Collaborate with research and infrastructure teams to deploy solutions at scale.
Seniority
Intern