AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
AI is analyzing your overall score…
Identifying your key strengths…
Evaluating your skill match against the job requirements…
Assessing your cultural and operational fit
Data and Algorithm Team
I am interested in solving problems by artificial intelligence, especially good at Artificial Intelligence in Operations & Management (AIOps). I have rich experience in artificial intelligence, data engineering, IT operations and software development, especially in neutral language process, machine/deep learning algorithms, distributed system and big data architecture, data center operations and site reliability engineering. Hands-on Specialties: - Algorithms: Decision Trees, SVM, SVR, SVD, KMeans, KNN, KD-Tree, PCA, DNN, CNN, GRU/LSTM, SGD, Q-Learning, Deep Reinforcement Learning - Statistics: Markov Process, Gaussian Process, Likelihood, and Monte Carlo - Neutral Language Processing (NLP): Gensim, Stanford NLP, and NLTK - Deep Learning and Machine Learning Kits: NumPy, SciPy, Skit-learn, XGBoost, Jupyter, Anaconda, PyTorch and Tensorflow with GPU supports, Tensorboard - Container Orchestration: Docker, Kubernetes and Rancher - AWS: EC2, S3, Lambda and Elastic Container Service - Data Processing: Kafka and Spark - Application Performance Monitoring (APM): Telegraf, cAdvisor, Prometheus, Grafana, and Chronograf; - Program Language: Python, C++, Java, Go lang, Html, JQuery, and Javascript - Frameworks: Spring (JAVA), Tornado (Python), Flask (Python) - Databases: MySQL, Redis, and InfluxDB - Multi-processes/threads coding: Python, Go lang and Java - Software Development Kits (SDK): PyCharm, IntelliJ, and GoLand - Load Balancer: Nginx and HAProxy - Linux Distributions: Ubuntu, Alpine, Debian - Agile Software Development Tools: Swagger, GitHub, ZenHub, Trello, Cloud9 and Slack
Northeastern University
Doctor of Philosophy (Ph.D.), Physics
January 1, 2007 – January 1, 2014
Peking University
Bachelor’s Degree, Physics
January 1, 2002 – January 1, 2006
Beijing No.4 High School
High School, general
January 1, 1999 – January 1, 2002
IntelliPro
Technical Partner
April 1, 2019 – Present
Beijing & San Francisco Bay Area · On-site
Huawei Technologies USA R&D Center (FutureWei)
Senior Data Engineer and Scientist
February 1, 2016 – April 1, 2019
Santa Clara, CA
CERN
Graduate Research Assistant
January 1, 2009 – August 1, 2015
Geneva Area, Switzerland
Northeastern University
Teaching Assistant
September 1, 2007 – December 1, 2008
Peking University
Research Assistant
July 1, 2006 – August 1, 2007
Job Candidate Matching
May 1, 2024 – Present
Project Leader and Main Developer Responsible for designing and writing architecture from scratch, and organizing cooperative development, deployment, publicity and maintenance work. • Collaborating with Columbia University, successfully tuned the job-talent match embedding model, achieving nDCG of more than 0.8 on Alibaba Cloud public data and about 0.5 on self-used data, and published related papers. • Based on the paper, applied model AI, combined with manual data cleaning, and tuned the core model, significantly improving the performance of self-used services, surpassing the original paper results. • Used distributed computing technology to compute and storage billions of data in very low cost • Optimized the Qdrant database so that the user is able to get talent recommendation in seconds from billions of data. • Researching reasoning RAG: Tuning the reasoning reranker model, providing detailed job matching reason analysis to help users make decisions. • Performance: half of the candidates recommended by AI have successfully got interviews with clients.
Resume Parser
May 1, 2019 – August 1, 2024
Project Leader and Key Developer Responsible for designing and writing the architecture from scratch, and coordinating development, deployment, and maintenance. • Led the development team and successfully developed a font-aware text tagging tool with tag review and proofreading capabilities, improving the accuracy and efficiency of data processing. • Designed and developed a comprehensive file reading framework compatible with formats such as PDF, Word, JPEG, and HTML, ensuring accurate parsing various document formats. • Improved the YOLO model to enable it to perceive the previous page's content, improving object detection and recognition accuracy by approximately 8%. • Improved the traditional BERT model to enable it to perceive font information, increasing keyword recognition accuracy by at least 5%. • Optimized parsing speed: Most resumes were parsed within 2 seconds, with recognition accuracy exceeding 95% for both Chinese, English, and PPT resumes. • Highly maintainable and extensible code. Through simple development, it has already implemented a variety of practical services, such as contract parsing and job description parsing, providing strong technical support for practical applications. • This feature has become one of the most commonly used in the daily work of group consultants, providing millions of structured talent data points to the group's database.
Autoshift
February 1, 2016 – December 1, 2017
This is a management software that automatically configures and maintains cloud service resources for applications. I am the main developer. • Invented a heuristic search algorithm that combines decision trees, reinforcement learning, and discrete stochastic gradient descent to search for optimal capacity planning. • Implemented Q-learning and regression tree reinforcement learning for online optimization of service instance counts and resource allocation. • Built a test application consisting of Nginx, Tomcat, Spring Framework, and MySQL, as well as a load test using Locaust, JMeter, and AWS Lambda. • Found the optimal capacity plan within 1.5 hours of the application's launch. After the customer application was deployed, the agent was able to respond within 1 minute, resolving latency and error rate spikes. Related Huawei Cloud patents are available. • Collaborated with the Berkeley Artificial Intelligence Research (BAIR) Lab to improve algorithms and research new scenarios, new requirements, and new solutions for reinforcement learning.
Cultural Fit Analysis
The candidate's diverse project portfolio, ranging from resume parsing and job matching to cloud resource optimization and high-energy physics data analysis, indicates adaptability and a broad technical interest. Their experience in both research (CERN, Huawei R&D, university collaborations) and industry (IntelliPro) suggests a blend of theoretical depth and practical application. The target role of ML Engineer aligns well with their core competencies in AI/ML model development, optimization, and deployment. The leadership experience also suggests a potential for mentorship and driving technical initiatives within a team.
Soft Skills & Operational Fit
The candidate demonstrates strong leadership and project management skills through their roles as 'Project Leader' and 'Technical Partner'. Their ability to coordinate development, deployment, and maintenance, along with advising multiple teams, suggests good operational fit. Collaboration with academic institutions (Columbia University, Berkeley AI Research Lab) indicates a proactive approach to research and development. The detailed project descriptions highlight problem-solving capabilities and a results-oriented mindset.