Intel® Student Ambassador for oneAPI
Overview
This program is targeted at undergraduate and graduate students who are passionate about technology and working with developer communities to promote learning, sharing, and collaboration. It provides students with a great opportunity to enhance their oneAPI skills and learn about the cutting-edge Intel® hardware and software products that drive the development of an open, standards-based programming model across multiple architectures to solve real-world problems.
Current Ambassadors
Adam Tuft, Durham University
Adam uses Intel® oneAPI toolkits to investigate the scheduling behavior of task-based parallel programming runtimes used in high-performance computing (HPC) applications. He particularly focuses on the efficient use of OpenMP* and DPC++ (the oneAPI implementation of SYCL*) for offloading to Intel® GPUs. Working within the ExCALIBUR task-parallelism research project, Adam uses the Intel oneAPI toolkits to develop Otter, a performance analysis tool. Otter aims to help developers port existing code to use task-based parallelism effectively.
Devesh Seethi, Northern Illinois University
Devesh builds multimodal frameworks on Intel® DevCloud using data from inertial sensors, vision, and audio to solve real-world problems to improve public health. He is creating a framework to track activity levels and behavioral patterns in Alzheimer patients using person reidentification, point of interest detection, motion tracking, and monocular depth estimation in the Intel® Distribution of OpenVINO™ toolkit. Adam is also exploring ways to fine-tune the framework for tracking activities of multiple patients simultaneously in different environments. Plus, his research aims to optimize the inference and computational stages of the framework with the help of Intel® Neural Compressor to make it compatible with resource-constrained edge devices and make the solution pervasive.
Harvey Johnson, University of Nottingham
Harvey is using oneAPI to implement an efficient and timely data preprocessing pipeline for AI. This system uses GPU and accelerator resources to compress and decompress hundreds of millions of data samples in seconds. This process simplifies loading large datasets for Gavin AI to allow for training on larger datasets without I/O or storage bottlenecks.
Poornima Nookala, Illinois Institute of Technology
Poornima's goals are to reduce overheads of tasking in parallel runtime systems as a step towards exascale computing. She and her team proposed XQueue, a novel lock-free multiple producer multiple consumer (MPMC), out-of-order queuing mechanism that can scale up to at least hundreds of threads. They built XTASK by integrating XQueue with OpenMP*, a widely used parallel programming interface for shared memory architectures. XTASK enables extremely fine-grained parallelism for native OpenMP applications that can run unmodified just by linking against their runtime library. Poornima uses several tools (like Intel® VTune™ Profiler, Intel® Trace Analyzer, Intel® Inspector, and others from the Intel oneAPI toolkit) in her projects for analyzing the performance bottlenecks and optimizing the runtime to scale up to hundreds of cores on modern many-core architectures.
Rachel Selina Rajarathnam, The University of Texas at Austin
Rachel works on methods and algorithms to accelerate the process of FPGA circuit placement for improving design scalability. Her DREAMPlaceFPGA, an accelerated global placement framework that's implemented using a deep-learning toolkit, uses CPUs and GPUs. The AI support in oneAPI provides a suitable platform for further development and extension of the DREAMPlaceFPGA by allowing the use of CPUs, GPUs and FPGAs in one platform.
Taisa Calvette, Instituto Militar de Engenharia
Taisa is developing a hybrid model with recurrent neural networks to estimate river flow in hydropower plants from pluviometer and satellite precipitation data. For this work, she uses the Intel® Deep Neural Network Library, Intel® Optimization for TensorFlow*, and Intel DevCloud.
Criteria to Become a Student Ambassador
- Enrolled in an undergraduate or graduate degree in any regionally accredited university
- Have a minimum of one year left until graduation
- Advocate for Intel® technologies on campus
- Share your work at events and conferences and host meetups on campus
- Create a project that describes the work or research you support on Intel® DevMesh
The Role of a oneAPI Student Ambassador
- Organize and deliver AI, high-performance computing (HPC), or visualization-focused training and workshops for students
- Evangelize Intel® Software Development Tools and resources
- Develop and share projects using Intel technologies that inspire fellow students
- Post blogs and articles to share what they're learning and best practices with peers
- Commit to act as an oneAPI Student Ambassador for one academic year and potentially spend another year as a mentor
Benefits
- Connect with Intel experts, other students, and professors working on oneAPI projects
- Extended access to Intel DevCloud
- Showcase oneAPI projects at industry and Intel sponsored events
- Support from Intel to organize local events such as watch parties and workshops (virtual or face to face) to build the oneAPI community
- Learn about the latest technology developments that are under a nondisclosure agreement (NDA)
- Get invited to exclusive events and training by Intel
- Earn recognition by Intel as an oneAPI expert
- Get reimbursed from Intel for participating in accepted conferences and events
- Apply for an internship at Intel