I build the systems that make ML work at scale — from silicon to software. At MetaMeta, my work spans AI accelerator pre-silicon validation (ASIC), ads model-hardware co-design, GPU kernel & compiler optimization, model training & serving, data infrastructure, and agentic AI systems.
Trained by Daniel AbadiDaniel Abadi to never trust a single point of failure (Ph.D., University of Maryland).
ML Serving, Inference & AI Hardware
Agentic kernel coding (KernelEvolveKernelEvolve, Meta Blog, Import AI, ISCA'26) · distributed serving · across NVIDIA · AMD · MTIA
@ Meta
Agentic kernel coding (KernelEvolveKernelEvolve, Meta Blog, Import AI, ISCA'26) · distributed serving · across NVIDIA · AMD · MTIA
@ Meta
Post-Training & Agentic RL
Post-training optimization · agentic RL infrastructure · data curation & synthesis · evals · experience graphs (TrellisTrellis)
@ Meta
Post-training optimization · agentic RL infrastructure · data curation & synthesis · evals · experience graphs (TrellisTrellis)
@ Meta
Data Infrastructure
KV stores · ML column stores (Bullion, CIDR'25Bullion) · file systems (FileScale, SoCC'23FileScale) · schema evolution (BullFrog, SIGMOD'21BullFrog) · HTAP · vector databases
@ Meta · ByteDance · Microsoft Research
KV stores · ML column stores (Bullion, CIDR'25Bullion) · file systems (FileScale, SoCC'23FileScale) · schema evolution (BullFrog, SIGMOD'21BullFrog) · HTAP · vector databases
@ Meta · ByteDance · Microsoft Research
Training Frameworks
Core contributor to PaddlePaddle (22k+ ⭐) distributed parallel deep learning training (LLM) framework
@ Baidu Research
Core contributor to PaddlePaddle (22k+ ⭐) distributed parallel deep learning training (LLM) framework
@ Baidu Research