• Storage Digital Platform Global Expansion: Guided the global expansion strategy of IBM’s first end-to-end storage product selling digital platform, enabling $31.4M opportunities creation and $3.2M orders completion in 2021, across 28+ countries in Europe, North America, and Asian Pacific. Developed a centralized dashboard and owned the metrics for evaluating the platform effectiveness across all stages of the marketing and sales funnel: campaigns, clicks, visits, responses, leads, opportunities, and orders. Streamlined multiple data sources from SalesLoft, Google Analytics, IBM Sales Cloud, and etc. • Sales Pipeline Diagnosis: Led the development of a diagnostic tool to guide global sales leaders’ target achievement strategy, which drove the immediate remediation of an 5-8% sales headwind in Q4 2020. The tool comprises early warning, performance benchmarking, and remediation suggestion, across 5 key dimensions: volume, value, win, quality, and resource. Implemented SuperImposition by Translation And Rotation (SITAR) model for estimating sales growth pattern, which proved to be 4.8x more reliable than conventional methods. Built an R Shiny UI to clearly deliver the tool’s insights to global sales leaders, which led to successful business adoption.• Cloud Paks Target Client Segmentation: Led the data analysis work-stream to promote the fast growth of IBM Cloud Paks, resulting in successful winning of $15M deals in 2020. Led science collaboration with a cross-functional team (demand generation, pricing, technology services, and enablement). Prioritized target clients based on client profiles, purchase history, and competitor information. Suggested to target existing growing clients, who are on the journey to the cloud with prior purchase of lead products and investment in the related domain.

• Digital Platform User Engagement: Invented a Reinforcement Learning agent that improves user engagement on digital platforms, demonstrated 3.5x performance improvement over naive interactions. The agent learns through users’ interactions on digital platforms and provides the optimal interventions to improve user engagement using Asynchronous Advantage Actor-Critic (A3C) algorithm with a multi-tier reward structure and a control layer. Proved the agent’s ability to improve usage and conversion rate in a simulated user environment using clustering analysis, distribution sampling, and mathematics transition matrix. Lead inventor of granted patent US11429869B2.• Hardware Storage Clients Retention: Provided prescriptive guidance to IBM hardware storage sales leaders on client retention and product cross-sale strategy, achieved 8.6% (7x baseline) opportunity creation rate in Latin America. Discovered the issue of dormant clients by building a dynamic visualization of client-brand relationship evolution. Analyzed client purchase cycle, expected value and cross-sale likelihood using Pareto/NBD model with Gamma Gamma Extension and Association Rule algorithm. Created action plans for the sales team based on analytical insights into which client to retain, when to reach out, which product to recommend, and the estimated value of the next purchase.• Data Science Community Build-up: Built a centralized science community platform. Drove data science awareness and adoption of broader tech community by piloting and lecturing ‘Intro to Data Science’ series. Mentored 6 data scientists to solve their domain problems with science models. Promoted Data Science among 4+ business units and nurtured further business collaborations.

• Drove North America zone office’s adoption of global data science assets by ingesting knowledge and experiences from Global Bud Analytics Lab.• Interlocked closely with directors of sales and marketing teams and influenced science-driven decision making.

• Order Suggestion: Launched Anheuser-Busch’s first B2B online ordering application with model-driven order suggestion functionality. Predicted customers’ future orders using the sales algorithm which utilized nested Clustering and LASSO Regression models with special treatment for weather, sports, and holidays. Deployed sales algorithm’s R and SQL scripts on Aster database to enable automated real-time order suggestion at the global scale. Conducted A/B testing using the APT platform to track and evaluate sales algorithm’s performance.• Customer Profiling: Created a seller tool for potential customer prioritization based on on-premise predicted sales potential and off-premise social impacts, which generated 3% top-line growth in NYC and rolled out to 13 major cities worldwide. Implemented Random Forest and Decision Tree algorithms to predict future customers’ potential volume even before sellers visit to customers. Performed Network Analysis to predict customers’ social influencing power for brand image establishment. Deployed the insights in a web-based user interface with integrated Google Map to facilitate sellers’ customer visits.

• Managed global tech sales team’s science initiatives and provided analytical consultancy to business stakeholders in multiple business units. • Recruited, mentored, and led 9 scientists to perform statistical analysis for the global tech sales team. Performed network analysis and visualization or social hub identification using Twitter, Yelp, and Foursquare API.• Built a text mining model to help the marketing team identify the optimum hyper-local events to sponsor.

• Evaluated an existing sales prediction model, identified low accuracy scenarios, and proposed an improvement to include influencing factors such as weather, sports, and holidays. • Web scripted Yelp and Foursquare data and built a Random Forest model to predict new customers’ volume potential.