New opportunity

Research Engineer, Pretraining

KiteFishAI · Gurugram, India

About this role

KiteFishAI is seeking a Research Engineer to build and improve Large Language Models from the ground up, working across model architecture, pretraining, data pipelines, and optimization. The role involves designing experiments, implementing research papers, and optimizing distributed training on GPUs.

Skills for this role

PythonPyTorchMachine LearningDeep LearningTransformer architecturesDistributed trainingGPU programmingCUDAData processingETL pipelinesNLPKubernetesFSDPDeepSpeedMegatron-LM