Leading a data science team takes more than the typical set of business management skills. You need specific know-how to articulate technology roadmaps, support a data-driven culture, and plan a data strategy that drives a competitive business plan. This book is written by practitioners for practitioners. If you are looking to manage your team better or work towards a seat at your company’s top leadership table, this book will show you how.

Berkeley China Summit was founded in 2016. It holds an annual summit and speaker series featuring high-level discussions about the latest discoveries, trends, and cross-border opportunities. Its mission is to encourage bilateral discussion and collaboration between the US and China.

Created and taught a graduate level course for credit and fraud risk modeling and mitigation at Tsinghua University Innovation: - Produced new course material to introduce Fintech and ML concepts to graduate studentsResearch: - Guided graduate level research in domain-specific feature extraction techniques

• Team: Hired 10+ data scientists in one month to fill a team capacity emergency, developed onboarding material to efficiently ramp batches of data science new hires• Planning: Drove quarterly planning process with strategic goals for embedding hiring marketplace dynamics into product definition, algorithm optimization, A/B test set up, and test result interpretations• Promoted LinkedIn’s brand of technical excellence and innovation in the industry by keynoting at two Talent Intelligence Experience conference in China to 500+ top talent acquisition executives

Transforming a micro investment app into a financial system for 5M+ investors through driving C-level data science initiativesInnovation: - Managing investor cash flow and investments with behavioral science- Establishing anti-fraud strategy and framework for a new debit card product preventing $2M/year in losses - Driving engagement-focused strategy for an acquisition-focused app with highly personalized user experiences by designing, developing, and launching a scalable and data-driven news feed content-creation infrastructure and a personalization framework that can generate $7M/year in revenue, achieving 50+% click-through rate - Fundamentally challenging status quo in direct deposit switch with a data-driven low-friction switch inventionTeam: - Recruited and assembled a tight and nimble team of seven data scientists- Designed a matrix organization to drive features in multiple product lines while developing core competencies- Created a candidate pipeline of 50+ inquiries/month by founding Meetup groups and speaking at conferences

Responsible for designing and driving C-level initiatives in data science for Yirendai, the first Chinese P2P Lending company to go public at NYSE. Reported directly to the CEO, built a data science team of 25:Innovation: - Filed two patents in fraud risk analytics- Completed 101 data science projects over 2 yearsRisk analysis:- Drove deep analytics social and e-commerce data and developed core signals in credit and risk model- Setup an organization and tool flow to continuously develop new signals- Stopped over US$30M of fraud cases a yearOperations:- Developed complete methodologies for analyzing user acquisition channels- Automated population stability monitoring channels with over US$50M spending per year- Setup customer tagging and experimentation infrastructure for over $300M of capitalTechnology:- Developed APP-SDK to report back user behavior for anti-fraud and customer insights- Introduced knowledge graph technologies based on Neo4j and drove adoption in multiple teamsTeam:- Built a team of 25 and trained six project leaders in data science

Constituted a culture of rigorous experimentation and analysis by building a data science team from scratch in the 150-employee job search engine that indexes over 10M jobs everyday, attracting 30M monthly unique visitors, serving 20K+ employers in 24 countries - Produced $1M/year in additional incremental revenue by designing and leading design/the development of the core auction algorithm for search engine based on learning-to-rank, a 6x ROI on the R&D efforts- Increased engineering velocity by designing and deploying a state-of-the-art A/B testing infrastructure- Constituted and drove new strategic projects including curating new large-scale golden data set from scratch- Advised Simply Hired executives on data-driven marketing strategies, sales enablement, business operations, and recruiting technology- Invited to speak at the White House in DC for job search technologies and policies using Big Data, as well as industry (HR Tech) and technology (Strata) conferences

Founding co-director of both centers focusing on application-driven research and teaching in computer science, to advance state-of-art in machine learning for human-centric computing, exploring core technologies such as speech recognition, NLP and computer vision- Defined research agenda and led teams of researchers to use the latest parallel computing technologies in machine learning and computational finance applications- Designed, developed, and taught one of the most popular grad-level course in parallel computing at CMU to more than 250 students on three campuses in two continents over three years

Vision: Use quantitative approaches to inform investment decisions.Execution: Architected a new quantitative analytics flow for investment risk analysis; led a team to develop the infrastructure to collect and explore data set, to discover and select features for analysis, and to implement the algorithms for risk factor analysis.

Helping organizations navigate the process and address the intricacies in significantly speeding up compute-intensive applicationsEvaluate industry-specific opportunities, mitigate deployment risk, and align acceleration technologies with clients' strategic roadmapsClients in computational finance and speech/video analysis fields have successfully transformed their businesses through increasing application performance by 14x to 750x using NVIDIA multi-core GPUs in CUDA and x86 CPU architectures using threads and SSE/AVX extensions

• Conducted and promoted computational finance research in the School of Engineering• Consulted for a leading global information company on utilizing highly parallel platforms • Published research paper on risk analysis in Super Computing Conference workshop

• Designed and delivered a course on parallel computation for 30 executives from Panama

Developed application frameworks in speech recognition and computational finance for application domain experts to productively and efficiently utilize multicore and manycore computing platformsResearch: Led multiple groups of 2-3 researchers, sometimes in inter-disciplinary teams, to produce 11 first-authored peer-reviewed papers, book chapters and patents.Education: Developed multiple sets of material in parallel computing and taught them in schools, client sites and 20+ seminars around the world.

• Accelerated deal closing by conducting competitive analysis, and presenting to global sales teams• Developed strategies for commercial software vendor facing an open source incumbent• Led discussions with Electric Cloud executives at Haas Business School through first-authored business case (https://drive.google.com/file/d/0B8Jn9MInY0K-cEg5TzJfWDNvZFE) Password: "business"

• 1 of 40 2-year Ph.D. fellowship recipients in the US (awarded in 2008-2009)• Invented novel techniques for high performance SIMD kernels (US Patent 2010/0082939)

• Designed logic and circuit of 3 microprocessors targeting high/mid-range server markets• First authored two US patents (7353364, 7426630) for key modules in flagship microprocessor