PhD Research Internships in Model Compression and Kernel Optimization at Meta Superintelligence Lab (Summer 2026) Meta in United States

Meta Superintelligence Lab's Kernels & Optimizations team is offering multiple PhD research internship positions for Summer 2026, located in Menlo Park, USA, and London, UK. The research focuses on advanced model compression techniques such as quantization, pruning, distillation, and sparsity to enhance training and inference efficiency. Interns will also work on architectural and kernel optimizations to reduce memory, compute, and latency while maintaining model quality, and unlock the full potential of GPUs for deep learning workloads.

Applicants must be senior PhD students expected to graduate in 2026 or 2027, with a strong publication record in top conferences like ICML, NeurIPS, ICLR, ISCA, ASPLOS, PLDI, SOSP, OSDI, or SC. The ideal candidate will have expertise in model compression, kernel optimization, and GPU-based training and inference. This paid internship provides an excellent opportunity for successful interns to transition to full-time employment at Meta after graduation.

To apply, candidates should submit their application via the Meta Careers portal. The application window is open for Summer 2026, with a deadline of October 1, 2026. For more information, refer to the provided links or contact Sandeep Parab, Engineering Manager at Meta Superintelligence Lab.