Statistician, Biostatistics & Research Branch, Epidemiology & Disease Control Division

• Oct 2021 - Present

Developed a bare-bones Markov model to project DM prevalence for target-setting Enhanced upon Markov model by leveraging multiple cross-sectional NPHS survey data and OR-based simplex algorithm to estimate transition probabilities for transitions between BMI-FPG states as inputs to Markov model to incorporate risk factor (BMI) into DM prevalence projectionDeveloped a script in STATA to interactively automate yearly aggregate data extraction from 2 complex data sources for sharing with external stakeholder - cut down turnaround time by approximately 12 folds. Validated successfully with prior years of extraction.

Intern, AI & Emerging Technologies, Data Analytics Division

• May 2021 - Sep 2021

Hospital Bill Estimation- Improved bill predictions by reducing bill underestimations relative to existing system by 2%, 10% and 16% based on 5%, 10% and 15% prediction intervals respectively- Validated solution and augmented patient experience with a host of ML interpretability

Graduate Student

• Aug 2020 - Sep 2021

Data Science Immersive Programme Graduate

• Feb 2020 - Apr 2020

Project 1: US College Exam Participation Analysis- Analyzed relationship between SAT and ACT participation rates using data visualization likescatterplots, heatmaps, boxplots; combined outside research to identify state of interest inpromoting SAT participation rates and tailored recommendations accordingly such asimplementation of SAT School Day programmeProject 2: Predicting Housing Sales Price- Analyzed relationships between housing features and sale price using data visualization likeheatmap, boxplots; built predictive models to predict sale price with a Kaggle RMSE of ~36,000,which is decent relative to the public leaderboard (20,000 – 300,000); R2 score on validationdataset was 0.88Project 3: Efficient Reddit Thread Filtering- Analyzed relative importance of various words to 2 distinct but similar Reddit threads withvectorizers and data visualizations like bar plots and word clouds; built predictive models topredict which thread posts belong to with an accuracy score of 0.96 on the validation dataset;ROC AUC ~1Project 4: Predicting West Nile Virus- Collaborated with classmates to analyze relationships between various variables using a variety ofdata visualizations like barplots, heatmaps, and distribution plots; built predictive models to modeloccurrence of West Nile Virus in Chicago from 2007 to 2013- Random Forest maximized detection of West Nile Virus (based on sensitivity) performed wellwith ROC AUC of 0.82 on the validation dataset; Kaggle submission ROC AUC was ~0.67Capstone Project: Local Coffee Outlet Recommender (based on data scraped from Yelp)- Built hybrid recommender to recommend local coffee-drinking places to users- Content-based Filtering-XGBoost, F1: 0.97; Model-based Collaborative Filtering-ALS, F1: 1.0- Managed to deploy a student-level coffee outlet recommender project on Heroku, based only on Content-based Filtering with XGBoost premised on few outlet features (Link: sg-coffee-recommender.herokuapp.com)

Executive (C&P unit under RDO)

• Aug 2015 - Feb 2020

-Manage joint research grant administration which involves liaising with various stakeholders from dissemination of grant publicity, invitation of reviewers, organization of grant shortlisting panel meetings, announcement of grant award results, to verification and processing of research grant claims-Conduct periodic reviews and updates of existing joint grant SOP to improve joint grant process-Create, update, and maintain grant budget and process tracking spreadsheets-Perform secretariat functions for joint research institutes-Coordinate component(s) of an annual health and biomedical congress such as poster exhibition-Draft research-related contracts to anchor research partnerships between parties-Provide regular reporting of KPIs to different stakeholders-Discharge adhoc duties eg. Analyzing company’s publication track record to churn out summary statistics and present in a digestible way to senior management to convince them to allocate funds for future research work

Ophthalmic Photographer

• Jul 2014 - Jan 2015

-Conduct a research study about retinal imaging in dementia-Acquire handling competencies of auto-refraction, fundus, and optical coherence tomography (OCT) machines-Measure choroidal thicknesses of about 330 human subjects from Alzheimer's Disease (AD), Vascular Dementia (VaD), Mild Cognitive Impairment (MCI) and healthy categories using a MATLAB program -Conduct a research study about retinal imaging in dementia-Acquire handling competencies of auto-refraction, fundus, and optical coherence tomography (OCT) machines-Measure choroidal thicknesses of about 330 human subjects from Alzheimer's Disease (AD), Vascular Dementia (VaD), Mild Cognitive Impairment (MCI) and healthy categories using a MATLAB program

Intern

• May 2013 - Aug 2013

Spearhead novel research into the possible solution of the global tetrameric structure of the tumor suppressor protein, p53. Experimented with different parameters in doing so and analysed the results via different fitness parameters and graphical plots. Primarily used martini coarsed-grained simulations for this purpose. Spearhead novel research into the possible solution of the global tetrameric structure of the tumor suppressor protein, p53. Experimented with different parameters in doing so and analysed the results via different fitness parameters and graphical plots. Primarily used martini coarsed-grained simulations for this purpose.