Prime Intellect logo

Applied Research - Continual Learning

AI
Prime IntellectLondonHybridFull-time3 weeks ago

Prime Intellect is hiring an Applied Research scientist for continual learning to build systems that turn agent trajectories, corrections, and failures into new tasks and model updates. The role involves comparing streaming RL, batched post-training, and distillation, and determining which improvements belong in model weights versus persistent memory, skills, prompts, and subagent state. Requires strong Python, hands-on experience with model-training or post-training loops, and strong experimental judgment.

What you'll do

  • Build continual learning systems that turn agent trajectories, corrections, and failures into new tasks and model updates
  • Compare streaming RL, batched post-training, and distillation
  • Determine which improvements belong in model weights or persistent memory, skills, prompts, and subagent state

What you need

Must-have

Continual LearningOnline LearningReinforcement LearningMeta-LearningMemory SystemsPost-TrainingAdaptive Production MLEvaluation

Soft skills

Experimental JudgmentFailure AttributionOwnership of Ambiguous ProblemsCommunicationWillingness to Change Direction
Applied ResearchContinual LearningReinforcement LearningAI InfrastructurePost-TrainingOpen SourcePythonLondonRemote