Building decentralized AI

- Built and managed the ML engineering team- Led the development of multiple products including ML workbench, a platform for graph ML andTigerGraph CoPilot, an AI toolkit for TigerGraph- Worked directly with potential customers to address their business problems such as fraud and churn- Collaborated with top researchers from academia and industry on graph ML

- Led a team of scientists and engineers to solve recommender systems problems for internal teams- Led research in reinforcement learning on graphs with application in explainable recommendations- Developed sequence-aware recommender systems which yielded .14% increase in annual revenue- Developed context-aware recommender systems which yielded .1% increase in annual revenue- Built Amazon’s knowledge graph of books and led research in graph embeddings

- Developed statistical unsupervised learning methods for high-dimensional network/graph data- Studied the stock market Flash Crash in US with high-frequency data and network methods- Developed hypergraph models and methods for novelty measures in science and technology- Led the Center for Network Analysis at Carolina

- Study higher-order interactions in complex systems using hypergraphs.- Infer public interest from online book co-purchase data.- Investigate the influences of conventions, biases and novelties in shaping the landscape of science.- Develop models for active learning of network interactions.- Investigate the innovating processes in US patents.

* Study theoretical models for coevolutionary networks with a focus on the evolving voter model;* Advance statistical algorithms for inferring implicit networks on which information diffuses;* Investigate diffusion-limited antibody-virus interactions in the human mucus layer;* Build network-based multiscale assessment tools for electrical properties of nano-composites; * Study the asymptotic behavior of high-aspect-ratio ellipsoids under shear flow.

Instructor for “Selected Topics in Mathematics”; undergraduate level course.

Teach students MATLAB and Mathematica for scientific computing. Instructor: M. Gregory Forest.

Instructor for "Pre-Calculus",Freshman-undergraduate level course

Developed efficient algorithms for contingency analysis of power grids via network theory and tools.

* Studied the Inner-City Asthma Air Pollution under the supervision of Rho, Inc. Chapel Hill, NC* Developed models for predicting indoor pollution from questionnaire and outdoor pollution data

* Developed a fast algorithm to compute loss distributions of portfolios using Gibbs sampler* Estimate the likelihood of serious losses of portfolios using importance sampling and milestoning

* Developed numerical simulations for the evolving voter model* Studied the mathematical structure of evolving voter model with two opinions

* Developed a clustering-based algorithm for deformations of large meshes with detail preserving* Designed an algorithm for rigid deformations of tetrahedral control meshes by Laplacian energy.