Joseph Kuo Email and Phone Number
Joseph Kuo work email
- Valid
Joseph Kuo personal email
Currently serving as a Deep Learning Software Engineer at Intel, I specialize in creating performance projections for deep learning models on future Intel Xeon CPUs. My past experience includes working as a research assistant at the Computational Imaging Science Laboratory, making open source contributions to the SciPy and Devito Python libraries. My expertise lies in machine learning, high-performance computing, and the intricate relationship between AI and hardware design.I thrive in environments that value collaboration, innovation, and the drive to push technological boundaries. Feel free to connect with me if you're interested in discussing machine learning, AI, deep learning, high-performance computing, or simply want to chat about the future of technology. I'm always open to meeting new people and exploring opportunities for collaboration.
-
Deep Learning Software EngineerIntel Corporation Jan 2023 - PresentSanta Clara, California, Us• Developed Xeon projections for large language models like GPT-J, ChatGLM, and LLaMA-2 70B, using Python.• Engineered optimization in tensor layouts, enhancing memory utilization in dense layers of large language models from 60% to 90% of STREAM bandwidth, leading to marked improvements in model performance for real-time token generation.• Collaborated closely with the hardware team to diagnose and resolve issues in MHA kernels and enhance uncore scaling.• Developed projections for TransformerLT performance on Intel CPUs, playing a pivotal role in the advancement of hardware-software co-design strategies. -
Research AssistantUniversity Of Illinois Urbana-Champaign Jan 2020 - Dec 2022Champaign, Il, Us• Enhanced transcranial photoacoustic image reconstruction efficiency, reducing computation times from 6 months to 15 minutes by implementing a supervised U-Net convolutional neural network.• Engineered a robust dataset comprising of 1,500 image volumes (768x768x384) for efficient machine learning algorithm training.• Optimized CUDA code to boost cache hit rate from 31.66% to 65.14% and successfully eradicated divergent branches, leading to a 4x decrease in runtime.• Improved parallel processing efficiency by 30%, going from 58% to 88%, through the strategic replacement of serial CPU MPI calls with asynchronous GPUDirect MPI calls.• Formulated analytic expressions for cylindrical acoustic wave scattering on an elastic cylinder, validating our numeric methods' achievement of linear convergence. -
Software EngineerUniversity Of Illinois Urbana-Champaign Jan 2019 - Dec 2019Champaign, Il, Us• Designed pedagogical tools to leverage virtual reality to teach Gauss’ Law, Ampere’s Law, and other topics inelectromagnetism• Reduced deployment time by ∼ 40 minutes to ∼ 5 minutes by setting up an CI system using Jenkins• Delegated tasks to a team of 20 and ran code review sessions
-
InternDistributed Technologies Research Jun 2019 - Aug 2019• Adopted ‘Nothing-At-Stake’ resource exhaustion attacks on Proof-of-Stake cryptocurrencies• Benchmarked attack on QTUM testnet, was able to fill the victim’s RAM at a rate of 2 MB/s RAM
Joseph Kuo Education Details
-
University Of Illinois Urbana-ChampaignElectrical & Computer Engineering -
University Of Illinois Urbana-ChampaignElectrical Engineering
Frequently Asked Questions about Joseph Kuo
What company does Joseph Kuo work for?
Joseph Kuo works for Intel Corporation
What is Joseph Kuo's role at the current company?
Joseph Kuo's current role is Deep Learning Software Engineer @ Intel.
What is Joseph Kuo's email address?
Joseph Kuo's email address is jo****@****tel.com
What schools did Joseph Kuo attend?
Joseph Kuo attended University Of Illinois Urbana-Champaign, University Of Illinois Urbana-Champaign.
Who are Joseph Kuo's colleagues?
Joseph Kuo's colleagues are Mohammad Kabir, Vinay Sindhur, Sant Ritu Awadh, Tareq Muhtaseb, Tanushree Vinayak, Greg Hunt, Walter Chiang.
Free Chrome Extension
Find emails, phones & company data instantly
Download 750 million emails and 100 million phone numbers
Access emails and phone numbers of over 750 million business users. Instantly download verified profiles using 20+ filters, including location, job title, company, function, and industry.
Start your free trial