GPGPU research expertise with direct CUDA kernel work (pinned grnn repo in Cuda language, co-authored 'GRNN: Low-Latency and Scalable RNN Inference on GPUs' 2019)
Deep GPU systems background via DeepSpeed-FastGen and ZeRO++ at Microsoft Research; now Researcher at OpenAI in SF
Seniority note: h-index 14 and 5+ years GPU work suggests mid-level rather than true junior
Hireability: MEDIUM — ~2 years into role at OpenAI (within typical transition window), no explicit openness signals, but tenure and prior job mobility (Microsoft → OpenAI) suggest willingness to move
XL
Xueshen Liu
medium hireability
Student Researcher@Google
Previously: Student Researcher @ Google
Ann Arbor, US
4th-year PhD at UMich (CS&E) focused on parallel computing and GPU/ML systems
Direct CUDA work: mm2-gb (GPU-accelerated DNA mapping), Foundry (CUDA graph materialization for fast LLM cold starts, C++), and HeterMoE (MoE training on heterogeneous GPUs)
Hireability: MEDIUM — 4th year PhD started 2022, likely 1-2 years from completion; wrapped up Google SRG internship Dec 2025 and returned to campus; website cv_update and position_update signals in Dec 2025 show career motion but not yet in prime transition window
ZY
Zhongming Yu
medium hireability
Student Researcher@Google
Previously: Machine Learning Engineer @ Intel
San Francisco, US
Strong GPU kernel engineer — papers on TorchSparse++ (sparse convolution on GPUs, 46 citations), SpMM heuristics on GPUs (31 citations), GeoT (segment reduction on GPU in CUDA/C++), and SGAP (sparse tensor algebra for GPU)
Pinned dgSPARSE-Lib CUDA repo
PhD at UCSD (year 4, started 2022), Student Researcher at Google since Apr 2025
Based in San Diego, US
Hireability: MEDIUM — year 4 PhD within typical transition window; current Google Student Researcher role is part-time/internship, not full conversion; no explicit job-seeking signals but no negative signals either