Applied Scientist - Trust and Safety (Multimodal Foundation Model) - Global Frontier Tech Recruitment Program - 2027 Start (PhD)
Core
Building multimodal foundation models and agentic moderation systems to protect users from negative content using state-of-the-art machine learning.
Role type
Senior IC Applied Scientist (Multimodal Foundation Models & Agentic Systems)
Builds
Large-scale MoE architecture training, RL-driven agentic decision-making systems, and multimodal safety foundation models for content moderation.
Domain
AI Safety, Content Moderation, Multimodal AI, Reinforcement Learning
Deliverable
production ML models
Required skills
PhD in CS/Data Science/AI, LLM research expertise, Python/Rust/C++ programming, deep learning frameworks (PyTorch, DeepSpeed, Megatron), distributed computing, RL, MoE, PEFT
Preferred skills
Published research papers, inference tuning and acceleration, GPU/AI accelerator expertise, LLM application & agent development evaluation
Technologies
PyTorch, DeepSpeed, Megatron, vLLM, GRPO, PPO, Langchain, GraphRAG, MCP
Responsibilities
Train and optimize large-scale sparse MoE architectures for cross-modal alignment; Develop RL-driven agentic systems for multi-step reasoning and tool collaboration; Engineer context assembly and multi-source evidence fusion strategies; Ensure generalization across 200+ languages and adversarial robustness against AIGC content.
Seniority
PhD level, hands-on IC
