- Led several research projects including a chatbot integrating OpenAI GPT-4, AWS, and heroku which demonstrated therapeutic alliance between participants and the chatbot therapist (see live on Telegram @CoachRecovery) - Machine learning using scikit-learn to identify individual difference predictors in two NIH-funded datasets. Data analysis of more than 500 fMRI sessions with python machine learning approaches including hyperparameter tuning, model selection, and cross-validation. - Successfully led data collection from 2,000 participants in a self-report survey with embedded A/B test design, identifying effective new ways to induce behavior change and reduce smoking craving along in at-risk vulnerable populations with relevant mechanisms of action - Published several research papers on AI describing balancing of multiple of objectives in RL agents - Designed a linear computational model of reward prediction error connecting neural activity to behavior and published the results in a leading neuroscience - Led a lunch discussion series on deep learning research and taught classes on Motivation and Emotion and Scientific Writing

Planning cognitive Evals for frontier Ai

For Auckland International Airports Limited: - Shaped government policy by presenting a Covid-19 model to the Ministry of health; recommendations we made were adopted within a month of presenting the model to their team - The model was built in R and presented in an ambitious RShiny data visualization web app with data from external sources, available at https://bnanalysis.shinyapps.io/border_covid_assessment/ - We published a paper about the science behind our model in the New Zealand Medical JournalFor SapienX/ESP: Time series modeling to identify electricity waste by technology in an industrial settingFor Litmaps. Optimizing data import pipelines for natural language corpuses

- Demonstrated For Sale By Owner prices are comparable to For Sale By Agent and identified Time-To-Sell as the key advantage for Agent sales, using explainable AI to carry out a linear regression market study on 100,000 property sales- Built an ML model using R for homeowner Propensity to Sell using user data and property information to predict propensity to sell where identified properties had 5x higher likelihood of selling vs. baseline; time since last sale was identified as a leading predictor- Designed a Supply and Demand PowerBI dashboard describing nationwide market supply and demand based on website engagement data, drawing interest from a major public housing developer interested in predicting demand- Improved Automated Valuation Model for homes based on a scikit-learn LightGBM model from a dataset of 500,000 properties; quantified model variance and wrote a report summarizing model accuracy- Productionized a deep-learning based model for computer vision to classify uploaded images and taught an internal class on computer vision for software engineers.- Created a Product Adoption PowerBI Dashboard to track adoption of a new product released- Created an automated monthly “Property Report” using RMarkdown to describe property industry activity

In my work at the University of Southern California I have:- researched risky decision-making through analyses of fMRI data written in python and R- designed and ran surveys assessing social psychological theory in morality, culture, and decision-making - wrote a wrapper R package to integrate caret machine learning with neuroscience packages - used python to build a computational model of motivation in decision-making recently published in the Journal of Research and Personality - taught 4 lab sections of 20-30 undergraduate students each for PSYC201: The Science of Happiness and PSYC 165: Drugs, Behavior, and Society.- Published my work in Social Cognitive Affective Neuroscience, Neuroscience and Biobehavioral Reviews, and other outletsExpecting graduation July 2018

- collected and analyzed longitudinal data on user behavior from over 1 million users using HIVE SQL log analysis data mining, describing patterns of product engagement over the course of a yearran surveys on several thousand users about their attitudes toward the product- shared insight and made recommendations about improvements to the products which informed a major product revamp in 2016-2017.

Working within the software services team to develop a .NET platform services website. I took the lead roll as the primary developer and system designer in building a new version of our customer support website.

I provided third-tier technical support to users of Vista’s software.

I taught laboratory classes of 20-30 students for PSYCH108: Individual, Social, and Applied Psychology.

I used MATLAB to test simplified algorithms for modelling neuron networks in the respiratory system; our work resulted in a publication listed below.

I managed social media and other tasks including statistical analysis and facebook application development.

I taught Oral English as a second language in Nanjing, China, to 6 classes of 30-50 students each.

I began as an intern through an Auckland University programme; the company was known as Proximity ID at the time. I continued working here part-time throughout my time studying. My contributions include technical design, implementation, and administration of the planetstream.tv website.

IT & Communications Manager, building a web site (PHP) and keeping company contacts up to date with company developments.

Customer support, C#.NE including HTML and JavaScript, Microsoft Access, and SQL Server administration.