Qiuwen Chen
Data AnalystResearch Engineer
Google DeepMind
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Research Engineer
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Avid learner
Syracuse University
Doctor of Philosophy (PhD), Electrical and Computer Engineering
January 1, 2012 – January 1, 2016
Beijing University of Posts and Telecommunications
Master of Engineering (MEng), Communications and Information Systems
January 1, 2009 – January 1, 2012
Beijing University of Posts and Telecommunications
Bachelor of Engineering (BEng), Telecommunications Engineering & Managements
January 1, 2005 – January 1, 2009
Queen Mary University of London
Bachelor of Engineering (BEng), Telecommunications Engineering & Managements
January 1, 2005 – January 1, 2009
Google DeepMind
Research Engineer
November 1, 2025 – Present
Mountain View, CA
Machine Learning Engineer
November 1, 2017 – November 1, 2025
Mountain View, CA
Microsoft
Software Engineer
October 1, 2016 – November 1, 2017
Redmond, WA
Microsoft
SWE Intern
May 1, 2015 – August 1, 2015
Redmond, WA
Syracuse University
Graduate Research Assistant
August 1, 2012 – October 1, 2016
Syracuse, NY
Alcatel-Lucent
Software Engineer Intern
January 1, 2012 – April 1, 2012
Beijing, China
Potevio Institute of Technology
Baseband R&D Engineer Intern
May 1, 2010 – January 1, 2011
Beijing, China
Beijing University of Posts and Telecommunications
Graduate Research Assistant
June 1, 2009 – January 1, 2012
Beijing, China
Real-time Anomalous Stream Detection using Spiking Neural Network
January 1, 2016 – August 1, 2016
- The first work to apply Neurosynaptic cores (IBM TrueNorth) to anomaly detection. - Provide real-time processing while consume power as low as a laser pen (50mW).
Fast Neural Network Text-Image Fragment Embedding
September 1, 2015 – March 1, 2016
- Design fragment padding and pooling for text-image alignment score calculation. - Improve training speed by 100X with theano implementation.
Bi-directional Mapping between Deep Image and Noisy Texts
September 1, 2015 – August 1, 2016
- Design cross-modal embedding layers that remove irrelevant text fragments. - Integrate deep face network and RCNN for rich context information. - Outperform state-of-the-art methods (Google and Stanford) on picture news dataset.
Defense against Training Set Manipulation based on Abnormal Concept Drift Detection
May 1, 2015 – Present
Machine learning techniques are developed at a skyrocketing rate in variety of applications, but this learning process is vulnerable from a security concern. Proposed attack methods as well-known as adversarial learning threat the accuracy of the learned model. Usually, data manipulation such as poisoning attack is a common attacking method to compromise a well-built learning model silently. To cope with this urgent situation, we proposed a general defensive method to identify the attacks based on abnormal concept drifting detection. Moving centroid change is used to indicate concept drifts, and both statistical regression-based detection and classification-based detection are applied to distinguish abnormal concept drifts from the normal ones. A case study for the proposed defensive mechanism is demonstrated and the evaluation shows that our method is effective in detecting the training set manipulation attack.
Parallel Confabulation Network Inference on GPU and Xeon Phi
July 1, 2014 – July 1, 2016
- Design efficient CUDA kernel algorithm, thread model and memory layout. - Significant Acceleration: 1000X faster response than single CPU implementation. - Scalability: Single card monitors 300+ traffic zones and 10K vehicles promptly.
Context-aware smartphone low power design
December 1, 2013 – Present
- Build prediction model for battery consumption patterns using user behavior. - Support control decision that save device power while maintaining QoS.
Self-structured Inference Network for Multi-feature Streaming Anomaly Detection
October 1, 2012 – May 1, 2016
- The first confabulation-based detection algorithm and high detection quality. - Reconfigurable software framework with quick model scripting
Intelligent Text Recognition System
August 1, 2012 – January 1, 2014
Inference-based sentence completion for machine reading and synthesizing.
Joint Transmission Scheme with Coding and Signal Space Diversity
July 1, 2009 – April 1, 2012
- Hybrid diversity (signal, code and space) techniques to aid wireless fading channels. - Implement baseband testbed on PicoChip many-core system. - The design is adopted as one of the standards in IEEE 802.11 (WiFi) specification.
Cultural Fit Analysis
The candidate's project portfolio is heavily skewed towards advanced research and development in machine learning, signal processing, and neuromorphic computing. While impressive, the projects are primarily academic or highly specialized R&D, which may not directly align with the typical day-to-day responsibilities of a 'Data Analyst' role that often requires more business-oriented data interpretation, dashboarding, and stakeholder communication. The candidate's experience at Google and Microsoft, particularly in roles like 'Machine Learning Engineer' and 'Research Engineer', indicates a strong fit for innovative, research-driven environments. However, the transition to a pure 'Data Analyst' role might require a shift in focus towards business impact and less on pure algorithmic innovation. The diversity of projects shows adaptability and a broad technical interest.
Soft Skills & Operational Fit
The candidate's project descriptions, while technically dense, lack explicit details on collaboration, leadership, or problem-solving methodologies beyond technical execution. The 'Tech Lead Manager' role at Google suggests leadership potential, but specific examples are missing. Operational fit cannot be fully assessed without behavioral data.