• Collaborated with a former Amazon Head of Marketing, Amazon GM of Data Science, and Microsoft product lead in launching a platform to provide data science services for organizations south of the Fortune 500 to meet a diverse set of business needs (funnel optimization, customer acquisition/retention, revenue forecasting, etc.).• Communicated findings to stakeholders ranging from senior science leaders to C-level execs. • Designed > 30 end-to-end ML models predicting population behavior, achieving > 80% accuracy with models in production using the programming languages (Python, SQL), popular ML frameworks (scikit-learn, gradient boosting, etc.), and cloud computing services (GCP, AWS).• Spearheaded the creation of recommendation models to optimize call center operations, which enhanced decision-making in targeted outreach strategies and reduced the total call costs ~20% based on the results of the randomized control trials. • Generated models predicting soil types for a U.S. County with only ~40% of county surveyed to understand the effects of pipe corrosion in different regions of the county. Presented the results to stakeholders via visualizations using geo-analyses and plotting packages (GeoPandas, Matplotlib). • Improved a data pipeline by applying distributed computing libraries and utilized table partitioning in data migration from the client’s CRMs (Salesforce) to an internal data warehouse (BigQuery), saved Storytellers > 90% in monthly cloud costs and had > 70% speedup in transport.• Built a Personally Identifiable Information (PII) detection and redaction tool for > 200,000 image and PDF files, preventing sensitive data transfer by leveraging services provided by AWS (Textract, PII Detector, SageMaker) and utilizing data from unstructured formats for model refinement.

• Used Monte Carlo methods to understand the nuclear quantum effects in molecules that undergo large amplitude motion (e.g. water clusters, hydrogen-ion clusters, etc).• Proposed a modified Monte Carlo approach to study quantum mechanics in water clusters which reduced the sample size by at least 20X. Run times decreased by 10X faster due to parallelization techniques. Resulted in 3 publications with > 30 citations and 4 presentations in 4 conferences. • Formulated an efficient sorting algorithm that classified millions of datapoints (> 1 TBs) into distinct structures, reducing the computational time by > 70%.

• Used quantum chemistry and molecular mechanics software to investigate the reactivity of the protein iso-penicillin-N-synthase and other related molecular systems and explored the geometric, environmental, and electronic structure factors responsible for the observed reaction dynamics. • Presented findings from studies at the ACS meeting in Dallas, TX, which resulted in a runner-up poster presentation.

• Led 2 general chemistry lab sections (24 students each), conducted office hours to reinforce course concepts, with ~10-15 students attending weekly. Received > 10 positive reviews.

• Determined various relationships between chromophores and the amount of waters that are added into the calculation using well-known computational chemistry software.

• Performed calculations of determining the reactivity of the Isopenicillin-N-Synthase (IPNS) and determined the reaction path and its energetics by looking at the reaction in various solvation conditions.