s
shar_asad1

Asad Ali

@shar_asad1

Physical Design Engineer

Pakistan
Engels, Urdu
Sommige informatie wordt in het Engels weergegeven.
Over mij
I am an Electronic Engineer specializing in GPU performance engineering, CUDA, PyTorch inference optimization, and AI hardware. I help AI teams reduce inference latency, improve throughput, optimize GPU memory, and increase utilization. My skills include Python, C/C++, CUDA, PyTorch, GPU profiling, mixed precision, torch.compile, and benchmarking. I built GPUOpt, which achieved 3.22x speedup and 68.97% lower median latency on ResNet-18 using an NVIDIA Tesla T4, with 100% Top-1 agreement on the tested batch.... Lees meer
s
shar_asad1
Asad Ali
offline • 
Gemiddelde reactietijd: 1 uur

Bekijk mijn diensten

AI-technologie advies
I will speed up pytorch cuda inference and optimize GPU performance
VanafUS$ 30 / projectMeer informatie

Portfolio

Werkervaring

Student

GPU/CUDA Optimization Engineer

Student

Apr 2026 - Present • 5 mos

I will Do GPU/CUDA Optimization for High and Low precision