Mar 2020 – May 2021 · 1 yr 2 mos

• Lead five talented Data Scientists, help to make LinkedIn the most trusted platform; oversee areas to defend and combat fake account, messaging abuse, scraping activities, and build better product features to foster member trust.

Mar 2019 – Mar 2020 · 1 yr

• Tech Lead in Growth Data Science Team, mainly focusing on Search Engine Optimization (SEO), guest acquisition, and network growth
• Led and co-chair the Spark cross-team forum to drive up Spark adoption across the Data Science organization

Jun 2017 – Mar 2019 · 1 yr 9 mos

• Lead Linkedin Learning Content Data Science team to empower intelligent Content decision, innovate across both the business front as well as the data science front to create leveragable results for the team. Two notable projects are:
• Initialized the Skill Match Index project, led three data scientists to develop the framework and algorithm to quantify student skill alignment toward the job industry, partnered with the insights and sales team on the go-to-marketing strategy; this project was adopted by the sales team and becomes a cornerstone of the higher education sector narrative to drive sales and customer relations.
• Initialized the Scala/Spark data foundation project, designed and developed architecture in Scala/Spark to address pain points on building a solid and reliable data foundation, collaborated with various data engineering teams to integrate LinkedIn state-of-the-art stack/tools. Such architecture emphasizes code modularization and reusability and is adopted as the next generation of data foundation across the data science org.
• (Fun activities) Led multiple teams to participate & win Hackdays, where ideas resulting in either roadmap change or patents
• 2017 Q3: “Best hack” in LinkedIn Learning Hackathon
• 2017 Q4: “Best Design” in LinkedIn Learning Hackathon
• 2017 Q4: “2nd price” in Data Hackathon (Sales solution theme)
• 2018 Q1: “2nd price” in Data Hackathon (Learning solution theme)
• 2018 Q3: Finalist (top 10) in Company Hackathon (Talent solution theme)
• 2018 Q4: Finalist (top 10) in Data Hackathon (GSO theme)
• 2020 Q1: Finalist (top 10) in Company Hackathon (Active community theme)

May 2016 – Jun 2017 · 1 yr 1 mo · San Francisco Bay Area

• Drive Growth and balance risk using Data Science
• carried out numerous A/B testing to power rider growth in China
• developed a price elasticity model to continuously optimize rider incentive distribution
• use analytics, A/B testing to power integration between fraud-prevention and vehicle dispatch system
• developed a collusion model to identify new fraud pattern ("know the unknowns")

Jul 2014 – May 2016 · 1 yr 10 mos · Greater Los Angeles Area

• Work in Research, Analytics & Development team, initializes projects and design algorithms to maximize company revenue generation, analyze product metrics to evaluate and optimize products, manage one direct report
• Product development:
• Led team on revenue-related project, designed innovative algorithms (using Baysian theory and discrete optimization) & developed a software package in Python to recover jeopardized sales, secured the company (estimated) $3M annual revenue
• Lead team to rebuild the core revenue system, architectured a generic machine learning framework, develope the system in Python/Spark language & Hadoop eco-environment, provide enhanced features and extendable functionality for better revenue characterization
• Initiated the "Trident" project as an architecture to optimize all recommendation engine related products on TrueCar website, lead team of six to develop Spark framework to consolidate data, generate insights, and design machine learning algorithms
• Insight exploration
• Initialized project to explore customer online behavior to study purchasing pattern and correlation with browsing experience, shed lights on several related projects to achieve improved prediction accuracy
• Carried out series analysis on in-stock vehicle recommendation performance, identified interactions between product features and lead performance, provide actionable recommendations
• Collaborated with executive team to identify potential revenue source, created a process to automatically screen out revenue-positive items, will contribute (estimated) $1M annual revenue

Aug 2013 – Jun 2014 · 10 mos · Greater Los Angeles Area

• Work in Research, Analytics & Development team, develop and implement algorithms for new products, leverage analytic insights to achieve effective marketing, build statistical models to predict vehicle pricing, and execute ad-hoc requests for various business needs
• Product related projects:
• Designed & developed popular configuration algorithm (patent pending, first inventor)
• Designed & developed vehicle recommendation engine (patent pending)
• Created an algorithm to recover comprehensive vehicle information (coded in Python), initiated as "The BuildX Project", provide enhanced data coverage for all related products (patent pending, first inventor)
• Marketing related projects:
• Data mining over keywords historic performance, identified a set of “hidden gem” keywords, increased SEM conversions/sales by 2~3% with minimal cost
• Managed SEM campaigns with consistently updating keyword/ad-copy/landing-page, analyzed A/B testing results to achieve better marketing performance

May 2009 – Aug 2013 · 4 yrs 3 mos · Durham, NC

• Independently finished six projects in Philip Handler Professor Weitao Yang’s group, published three first-author peer-reviewed papers on top journals (others under preparation), and presented twice at national conferences
• Algorithm design projects
• Devised an innovative algorithm and programmed 5,000+ lines code in C, accelerated ten-fold computational speed for solvation free energy simulation
• Created the first technique to quantitatively characterize non-covalent interactions in biological system, developed a stand-alone software to evaluate protein-ligand interactions
• Data analysis projects
• Carried out Monte Carlo and Molecular Dynamics simulations, analyzed hundreds of gigabytes simulation data with Perl and Linux Shell programming to study enzymatic reaction mechanisms
• Evaluated the cross relationship between proton and electron transfer by simulations, and provided the first theoretical evidence on how proton-electron transfer pathway is altered by electrode potential
• Explored the Quantitative structure-activity relationship between an innovative descriptor and molecule solubility, and automated the regression fitting process with Perl and R programming

Aug 2008 – May 2009 · 9 mos · Durham, NC

• • Lectured recitation class for 50+ undergrad students on general chemistry

Jun 2007 – Sep 2007 · 3 mos · Hefei, Anhui, P.R.China

• Investigated alloy magnetic properties using computational simulations, and published a first-author paper
• Advisor: Jinlong Yang

Sep 2006 – Mar 2008 · 1 yr 6 mos

• Managed four independent contest projects with teams consisting of three students each, conducted hypothesis-driven approach to build the mathematical model, and resulted in three national awards and one regional award
• Developed a transportation search engine based on sparse connection matrix, generated optimal routes based on starting point, destination, and method of locomotion (foot, bus, or train) in complex transportation system
• Invented a “virtual world simulation” model to evaluate each country’s health care system and identified that disease prevention is the key factor to improve the current system performance in U.S.
• Designed a revamped U.S. kidney transplant system, which decreased patient waiting time from 3.0 to 1.1 years
• Conducted risk portfolio analysis to optimize company resource allocation and projected a 30% boosted profit