Kunal Vaishnavi
TRS at Acetechnologies
San Jose, California, United States
Summary
🤩
Rockstar🎓
Top SchoolKunal Vaishnavi is a software engineer based in San Jose with three years of hands-on experience focused on machine learning engineering and performance optimization. At Acetechnologies he works as a TRS, bringing practical expertise in integrating and benchmarking ML frameworks—most notably adding PyTorch 2.0 support to Microsoft’s high-profile ONNX Runtime transformer benchmarks. His contributions include tuning attention mask handling and introducing a temperature parameter for Whisper model workloads, reflecting attention to both correctness and inference performance. Comfortable working in open-source ecosystems, he bridges research-grade models and production inference tooling. Trained at Panjab University, he combines a strong engineering foundation with a bias for measurable performance improvements in ML pipelines.
3 years of coding experience
Panjab University
Github Skills (7)
pytorch10
machine-learning10
benchmarking10
benchmark10
onnxruntime10
python10
deep-learning9
Programming languages (2)
Github contributions (5)
microsoft/onnxruntime
Jan 2023 - Jan 2023
ONNX Runtime: cross-platform, high performance ML inferencing and training accelerator
Role in this project:
ML Engineer
ML EngineerContributions:529 reviews, 1 commit, 81 PRs in 1 day
Contributions summary:Kunal primarily contributed to the integration and benchmarking of PyTorch 2.0 within the ONNX Runtime (ORT) transformer benchmarking framework. They added PyTorch 2.0 as an option for running transformer benchmarks, modifying the benchmarking script to include it, and updating the shell script used to run the benchmarks. They also made changes related to the handling of attention masks and the addition of a temperature parameter, indicating a focus on optimizing the performance and functionality of the Whisper model.
runtimetrainingtensorflowai-frameworkaccelerator
kunal-vaishnavi/transformers
May 2023 - Oct 2024
🤗 Transformers: State-of-the-art Machine Learning for Pytorch, TensorFlow, and JAX.
Contributions:23 pushes, 2 branches in 1 year 5 months
Find and Hire Top DevelopersWe’ve analyzed the programming source code of over 60 million software developers on GitHub and scored them by 50,000 skills. Sign-up on Prog,AI to search for software developers.