Research Scientist: Safer Frontiers in Agent RL & AI Safety

ByteDance; San Jose (CA); On-site; USD 254,000 - 588,000; Full time

Benefits offered by this job

Health insurance

401(k) with company match

Paid parental leave

Disability coverage

Job summary

ByteDance Seed seeks researchers to advance frontier models and agents with a strong emphasis on agent reinforcement learning, safety, and controllability in tool use, code execution, and long-horizon tasks. The role collaborates with safety evaluation and preparedness teams to translate research into safe, reliable AI systems that benefit people.

Applicants should have deep ML knowledge, Python and PyTorch proficiency, and hands-on experience in agent RL or related areas.

Qualifications

• Strong foundations in ML, deep learning, and statistics.
• Proficiency in Python and PyTorch with ability to implement algorithms and run experiments.
• Hands-on experience in agent RL or model training, dataset development, or ML infra.
• Ability to formulate testable hypotheses and communicate results clearly.
• Commitment to safety alongside capabilities in ambiguous problems.

Responsibilities

• Agent RL & safety training: design environments, rewards, and data to reduce deception and ensure safe autonomy.
• Frontier risk assessment: evaluate dangerous capabilities and model threats with domain experts.
• Automated red teaming: build automated attack generation and adversarial training to harden models.
• Monitoring & oversight: develop oversight models and measure effectiveness and costs.
• Failure analysis & deployment: convert real-world failures into reproducible experiments and pinpoint root causes.

Skills

ML foundations

Python

PyTorch

Agent RL

Experiment design

Tools

Python

PyTorch

Job description

ByteDance Seed seeks researchers to advance frontier models and agents with a strong emphasis on agent reinforcement learning, safety, and controllability in tool use, code execution, and long-horizon tasks. The role collaborates with safety evaluation and preparedness teams to translate research into safe, reliable AI systems that benefit people.

Applicants should have deep ML knowledge, Python and PyTorch proficiency, and hands-on experience in agent RL or related areas.