About the Role
Desplácese hacia abajo para ver todos los requisitos del puesto y las responsabilidades que pueden esperar los candidatos seleccionados.
Senior Post‑training & Efficiency Engineer (RE3) – responsible for leading the technical optimization of advanced post‑training pipelines including DPO, GSPO, DAPO, CISPO, and other preference‑alignment variants. Design and implement efficiency strategies such as model quantization and memory‑efficient training to maximize the use of HPC resources. Supervise experimental workflows, coordinate model deployment, and mentor junior team members.
Key Duties
Lead the technical optimization of advanced post‑training pipelines including Direct Preference Optimization (DPO), GSPO, DAPO, CISPO and other modern preference alignment variants.
Design and implement efficiency strategies such as model quantization and memory‑efficient training to maximize the use of HPC resources.
Supervise and refine experimental workflo...