AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Senior Staff ML Engineer & Manager @ Qualcomm | On-Device Agentic AI · VLA · Multimodal Perception
I lead the development of on-device agentic AI pipelines at Qualcomm — perception, reasoning, planning, and tool use — deployed across smart glasses, AI PCs, and agentic IoT systems. My team builds the full stack: multimodal perception (segmentation, detection, depth), VLM/LLM reasoning, retrieval and memory, and efficient on-device inference under tight latency and power budgets. Over 8+ years at Qualcomm, I've shipped flagship computer-vision features that reach millions of devices. Limitless Segmentation on Snapdragon 8 Elite, our Segmentation-based Cognitive ISP (winner of the 2023 Edge AI & Vision Product of the Year Award), and our personalized multimodal AI for smart glasses are featured across Qualcomm's developer ecosystem and major industry events including MWC, CES, and CVPR. I'm the leading and primary organizer of the IEEE Low Power Computer Vision Challenge (LPCVC) and the CVPR Workshop on Efficient Deep Learning for Computer Vision (ECV) for both 2025 and 2026, where I'm helping Qualcomm build an edge-AI developer ecosystem now exceeding 3,000+ members. LPCVC 2025 brought together 77 teams from 25 countries — featured in the IEEE Computer Society's interview report. My research spans real-time segmentation, model optimization, efficient transformer architectures, and most recently agentic multimodal systems. I've authored 24+ papers in top venues including ICML, AAAI, ICLR, KDD, IJCAI, and SIAM journals — with my KDD 2020 paper AutoShuffleNet selected as a Most Influential KDD Paper — and hold 18 granted patents in computer vision and on-device AI. I earned my Ph.D. in Mathematics from UC Irvine under Prof. Jack Xin, with a foundation in numerical optimization and compressed sensing. Areas I'm most excited about right now: · On-device agentic AI and VLA (Vision-Language-Action) systems · Multimodal perception for embodied AI — rob
UC Irvine
Doctor of Philosophy (PhD), Applied Mathematics
January 1, 2012 – June 1, 2017
Shandong University
Master's degree, Computational and Applied Mathematics
January 1, 2009 – January 1, 2012
Shandong University
Bachelor of Science - BS, Applied Mathematics
January 1, 2005 – January 1, 2009
Qualcomm
Senior Staff Engineer (Machine Learning), Manager
November 1, 2024 – Present
Qualcomm
Staff Engineer (Machine Learning)
November 1, 2020 – November 1, 2024
Qualcomm
Senior Engineer in Computer Vision and Deep Learning on Mobile Devices
July 1, 2017 – November 1, 2020
Zillow Group
Data Science Intern
June 1, 2016 – September 1, 2016
Seattle, Washington
Black Knight Financial Services
Data Scientist Intern
June 1, 2015 – November 1, 2015
Irvine, CA
UC Irvine
Graduate Student Researcher
September 1, 2012 – June 1, 2017
Orange County, California Area
Shandong University
Graduate Student Researcher
September 1, 2009 – July 1, 2012
Jinan, Shandong, China
Cultural Fit Analysis
The candidate's background is heavily skewed towards Machine Learning, Computer Vision, and Deep Learning, with a strong focus on on-device AI and engineering leadership. While there is early experience in data science (internships), the primary career trajectory is not directly aligned with a pure 'Data Analyst' role, which typically emphasizes data manipulation, visualization, statistical analysis, and business intelligence rather than advanced ML model development and deployment. The breadth of skills is deep within ML/AI but less diverse in traditional data analysis tools and methodologies. This suggests a potential mismatch with the target role's core responsibilities.
Soft Skills & Operational Fit
The candidate's experience at Qualcomm, progressing from Senior Engineer to Senior Staff Engineer (Machine Learning) and Manager, suggests strong leadership, project management, and problem-solving skills in a demanding technical environment. The descriptions imply an ability to drive projects from research to product, indicating a results-oriented approach. However, specific soft skill assessments are not available.