- Design, build, and deploy LLM-powered agents to detect and address spam, harassment, and other policy-violating content.
- Develop agentic workflows to automate complex tasks in content and user moderation systems.
- Design and improve machine learning systems that adapt to evolving abuse patterns and new challenges.
- Evaluate agent performance through offline evaluations, online experiments, and production monitoring.
- Balance accuracy, reliability, latency, and cost when evaluating system performance.
- Investigate failure cases, experiment with new approaches, and iterate on solutions.
- Identify opportunities to apply machine learning and LLMs to content quality and safety.
- Collaborate with experienced engineers and own projects from early development through production.
PythonKubernetesMachine Learning+4 more