AI Research Engineer

Company:

Qualcomm Technologies, Inc.

Engineering Group, Engineering Group >
Machine Learning Engineering

A leading technology innovator, Qualcomm pushes the boundaries of what’s possible to enable next‑generation experiences and drives digital transformation to help create a smarter, connected future for all. Our team of engineers and computer scientists work on projects related to high‑performance embedded and edge computing, advanced numerical / scientific computing and algorithms, state‑of‑the‑art optimizations, LLM and vision‑based model accelerations‑device performance optimizations, and quantization techniques, supporting internal and external clients at/of Qualcomm.

Our research is at the cutting edge of AI R&D, is highly impactful with products and customer applications, and is making relevant and high impact both in research realm and real‑world customers.

As a Machine Learning Engineer, you will work on a range of projects within a small technical team, conducting fundamental research that creates innovative machine learning methodology and achieves beyond state‑of‑the‑art performance. You will be closely guided and mentored by one of our experienced researchers, providing you with an unparalleled learning experience. You will have the opportunity to publish your own research papers, and to attend client meetings and academic conferences.

We support Qualcomm’s mission for making the world a better place one wireless connection at a time, making AI smarter, cars drive better, networks move data faster, and extending reality.

By joining our team, you can help Qualcomm Win Together, Achieve Excellence, Make the Impossible Inevitable, and Do the Right Thing.

• Cutting‑edge compiler optimizations for new industry‑leading Qualcomm AI chips.
• New approaches to and applications of nonconvex optimization.
• Developing new algorithmic implementations for Qualcomm hardware, grounded in mathematical and linear algebra theory and practice, improving on state‑of‑the‑art standard approaches.
• New approaches to accelerating inference and learning via algorithm, compiler and hardware innovations.
• Drop‑in replacements for traditional attention‑based transformer architectures.
• State‑of‑the‑art R&D in LLM inference efficiency algorithms, efficient model architecture design, and LLM training.
• Developing creative solutions with consideration of practical challenges on device.
• Implementation and evaluation of possible solutions in both simulation and on‑device environments for GPUs, Qualcomm NPUs, and heterogeneous computing environments for image recognition, autonomous driving, and general vision‑based environments.

• An ideal candidate will have:
• A solid intellectual and cognitive ability, motivation, a strong history of achievement, and an ability to synthesize knowledge from a variety of sources.
• A strong math foundation (theory and applied), numerical computing background, programming and engineering skills, as well as excellent oral and written communication skills.
• A firm math background, as well as a strong knowledge of advanced algorithms.
• A passion for Computer Science research and application.
• A genuine interest in the spectrum of Qualcomm’s 5+G, AI and other distributed intelligence projects.
• The flexibility to move among diverse sets of research topics.

• Research experience, including publication history – AI R&D preferred.
• Background and understanding of the mathematical foundations of deep learning.
• Strong programming / engineering skills (experience with version control, Git Hub, etc).

• PhD in Computer Science, Applied Mathematics, Electrical Engineering, or related field.
• Experience in LLM efficiency research such as efficient attention, inference acceleration, or KV cache compression.
• Deep experience with Pytorch, strong understanding of Python and packages.
• Experien…