Data Scientist

Kevin Xuan

*************@*****.*** 408-***-**** https://www.linkedin.com/in/kevin-xuan https://github.com/kevinxuan SUMMARY

Data scientist with 4+ years of experience in

● Delivering understanding and analytics work on Advertiser resolution path journey and Facebook user story consumption

● Creating detailed data analytics documentation and presented data results and visualization to all levels of audiences

● Communicating frequently with stakeholders, interdisciplinary teams, and clients to develop and enhance products

● Building, implementing, and documenting end-to-end machine learning data pipelines and solutions SKILLS

● Python, SQL, R, Linux/Unix, Docker, Spark, AWS, Git, Tableau, Scikit-learn, Numpy, Pandas, PyTorch, Matplotlib EDUCATION

University of California, Berkeley December 2022

● Master of Information and Data Science GPA: 3.96 Coursework Included: Applied Machine-Learning, Data Structure, Data Engineering, Data Visualization, Machine Learning on Large Scale, Natural Language Processing (NLP), Object Oriented Programming, Statistics, System Engineering University of California, Davis June 2017

● Bachelor of Science in Mathematical and Scientific Computation GPA: 3.53 EXPERIENCE

Meta (Contractor)Menlo Park, CA

Data Scientist January 2023 - Present

● Led analytic project evaluating understanding of user resolution path in Advertiser account issues, formulated detailed findings and recommendations, and collaborated with cross-functional teams to implement recommendations into production

● Conducted opportunity sizing on enabling sound on for users viewing Facebook music stories, made product launch decision based on designed experimentation results, and increased user story consumption by 5% and production by 2%

● Designed and executed A/B testing, define key performance, ecosystems, and guardrail metrics to evaluate feature impact to users, root caused regressed metrics, and informed business decisions on product feature launches

● Created visual data dashboard on top level data metrics, compiled codes to validate metric calculations, and performed code reviews on data tables via SQL and Python, reducing ~10% coding time for stakeholders and leadership team UniDT Technology Shanghai, China

Machine Learning Engineer October 2020 – December 2022

● Constructed object detection and pose estimation AI model to search human figures inside a video and utilized feature encoding to extract time frame intervals which these figures appear, achieving an accuracy of 95% in human detection

● Optimized model inference time by 50% and GPU memory usage by 75% via applying optimized engineering implementation methods such as multithreading and switching deep learning PyTorch models to TensorRT type models

● Designed, compiled, and optimized high-quality code in Python to construct scalable API services, utilizing Docker for deployment in production environments, and ensuring robustness, scalability, and compliance with industry standards R2.AI San Ramon, CA

Data Scientist August 2017 – October 2020

● Contributed to different phases of AutoML platform R&D by constructing, implementing, and troubleshooting modules: data quality check, data cleansing, data analysis, feature transformation, algorithm development, and model evaluation

● Worked closely with stakeholders to derive business demands and requirements, extracted tabular data from database via SQL, built ML predictive models via Scikit-learn, and delivered model evaluation and data visualization via Tableau PROJECT

Advertiser Opt Out Sizing (Meta)

● Conducted opportunity sizing analysis for advertisers expressing interest in opting out of the Ads sales program, created rule engine to distinguish advertisers strongly insisting on opting out, compiled detailed analysis encompassing revenue loss of Advertiser opt-outs, and presented business recommendations to senior leadership team on how to recover advertiser churn

Contact this candidate