Oversee data science research and activities in drug discovery and preclinical development:o Automated Dose Prediction (ADP): efficient and accurate compound triaging for 30+ Merck discovery programs with high-throughput human dose predictiono Unified Human Translation: A universal computational framework for human translationo Dynamic ROP: algorithmic decision-making to rival human experts in research operation plan (ROP) by translating predictivity into actionable decisionso Conformal Compound Selection: algorithm that enables efficient & accurate compound screeningo AlphaPK: generative model for pharmacokinetics/toxicokinetic (TK) curveso GenAI Report QC: streamlining study QC and report writing using large language models (LLM)o Controllable Molecular Generation: medicinal-chemistry aware molecular generative modelo QSAR-Complete: adaptive approximation towards human ADME endpoints through data completiono Bayesian IFS: Iterative Focused Screening of large compound library via scalable Gaussian process

Champion Data and AI/ML strategies in Drug Discovery and Preclinical Development:o Led the AI/ML vision and implemented strategic AI/ML initiatives for preclinical development.o Built a high-performing data science team that excelled in delivering impactful solutions.o Seamless development and deployment of cutting-edge AI/ML solutions to support pipeline operations, enhancing organizational efficiency.o Drive impactful internal and external collaborations with strategic partners, establishing valuable networks that enrich the organization's research capabilities and fostered innovation.

● 2021 STAT Wunderkinds for leading data science vision in drug R&D● Project representative DMPK- Investigate, Interpret, and communicate DMPK, safety, and efficacy data to the drug discovery project team to navigate effective drug design and risk mitigation- Develop translational PK/PD strategies and human dose prediction/simulation for CNS programs● Lead the build of data science capabilities in preclinical development- ROP coPilot: AI agent to augment decision-making in research operation plan in drug R&D;- Coaches data scientists and lead data science education across preclinical development;- Lead the development of automatic adaptive dose prediction through discovery to preclinical candidate;- Lead the development of generic predictive framework for all preclinical assays- DeepRatIV: rat IV PK curve prediction and human dose simulation directly from compound structures- Lead the build of cloud-based IT infrastructure and data architecture for preclinical development- Build integrated data visualization platform for DMPK and all CNS programs

● Early Talent Development Program: cross-functional training in Clinical Pharmacology, DMPK, Safety Assessment, Clinical Supply, Formulation, and Digital Health● Digital Analytics and Technologies, 2021Data modeling on WATCH-PD, clinical assessments on Parkinson's disease with new digital measures including an Apple Watch, an iPhone, and a full-body sensor system● Safety Assessment, 2020Predictive modeling on clinical drug induced liver injury (DILI)● Pharmaceutical Sciences & Clinical Supplies, 2019- Formulations development for enhanced oral bioavailability;- Investigated the impact of spray drying conditions to tablet compatibility- Designed data pipeline for Merck’s next-generation manufacturing facility (FLEx Center) through Amazon Web Service (AWS); - Developed time-series analysis tools to extract key material features (e.g. viscoelasticity, plastic deformation energy, elastic recovery energy etc.) during tablet compaction● Quantitative Pharmacokinetics and Pharmacodynamics, 2018- Populational PK by NONMEM; - PK/PD modeling and simulation; - Phase I dose selection, and non-compartmental analysis; - Led the design and development of machine learning methods for Exposure-Response modeling and simulation to predict the Phase 2 and Phase 3 outcome of combination therapy of anti-HCV drugs; - Supported simulation, modeling, and QC for the sNDA submission of Zerbaxa®● Department of Pharmacokinetics, Pharmacodynamics and Drug Metabolism, 2017- Enabled oral delivery of therapeutic peptides through DMPK characterization (enabled IND filing of MK-0616)- Advanced Deep Learning model to predict the cross-collisional section (CCS) for isomer differentiation through ion mobility mass spectrometry- In-licensing assessments, due diligence, and de-risking on peptide and protein therapeutics

• Led analytical method development and lead-optimization for large molecule therapeutics• Invented and patented 2 algorithms, line-graph and cut-vertex algorithms, for metabolic soft-spot identification of complex therapeutic peptides and proteins (invited for talk in the International Consortium for Innovation & Quality in Pharmaceutical Development)• Authored, reviewed, and filed INDs for peptide and protein therapeutics• ADME due-diligence on in-licensed large molecule therapeutics• Led, coached and mentored junior scientists

• Designed algorithm for ProSight Lite software for glycoprotein top-down MS/MS analysis by integrating Unicarb and Uniprot database (Proteomics.15(7):1235-8).

• Designed an analytical platform to characterize low-level isomeric impurities and degradants in therapeutic proteins formed during formulation. (Anal. Chem. 85(19):8964-7)• Elucidated protein isomers through Electron Induced Dissociation (EID) to. (Anal. Chem. 88(11):5914-9))