About this role
Research and develop post-training methods (e.g., supervised fine-tuning, parameter-efficient fine-tuning, and reinforcement learning) for large multimodal transformer models and build evaluation, experimentation, and inference infrastructure to support deployment in drug discovery workflows.
Skills for this role
PythonPyTorchDeep learningTransformer modelsTraining large-scale transformer modelsReinforcement LearningRLHFRLAIFPPOGRPOReward-based optimizationSupervised fine-tuningFull-parameter fine-tuningLoRADPOHyperparameter optimizationOptunaRay TuneExperiment trackingWeights & BiasesDockerCUDAKubernetesInference optimizationMixed precisionKernel optimizationQuantizationDistributed trainingHPCBenchmarking and evaluation frameworks''Reproducible experimentation''Clean 코드