Research Engineer, AI Agents for Creation Tools

TikTok
San Jose, California

About the job

We are building the next generation of AI-native creation tools powered by intelligent AI agents. Our mission is to transform how people create by embedding agentic AI directly into creative workflows—from AI-powered camera effect generation to AI-driven video editing and storytelling. Our team operates at the intersection of AI research, systems engineering, and creative product development, translating advances in agentic AI into scalable, real-world creation tools used by millions of creators. We work closely with product and design partners to ensure that sophisticated AI systems result in simple, powerful, and intuitive creative experiences.

Responsibilities

Research and develop agentic AI systems for creative tools

Design agent architectures for planning, reasoning, memory, and multimodal interaction

Build training, evaluation, benchmarking, and experimentation pipelines for agent systems

Integrate large language models and generative models into end-to-end agent workflows

Prototype new agent capabilities and validate them in real user scenarios

Collaborate with engineering teams to productionize research-driven systems

Qualifications

Minimum

Bachelor’s or Master’s degree in Computer Science or a related field

3+ years of related industry experience

Strong background in AI agents, LLM systems, or multimodal AI

Experience with model training, evaluation, and experimentation workflows

Strong programming skills in Python, C++, or similar languages

Ability to translate research ideas into practical systems

Preferred

Research or applied experience with tool-use agents, coding agents, or autonomous task-execution agents

Experience studying or improving agent behaviors such as tool selection, planning with tools, code-based execution, or error recovery

Experience applying AI to creative domains such as camera effects, content creation, or video

Publications or applied research contributions in agentic or generative systems