I majored in Data Science and Statistics, and minored in Economics at Minerva Schools at the Keck Graduate Institute. I traveled, worked, and studied in seven different countries around the world 🇺🇸 🇰🇷 🇮🇳 🇩🇪 🇦🇷 🇬🇧 🇹🇼 throughout my college education. Since I was young, I always enjoyed solving problems and building things. Writing software gives me the same kind of enjoyment I had when I built my own catapult as a nine year old kid.

Passions: Computer Science, Mathematics, Physics, Robotics, Economics, Finance

1. Causal Inference
2. Machine Learning
3. Bayesian Inference
4. Computational Modeling
5. Econometrics

Using statistics to prove empirically a causal relationship, and calculating the true treatment effect.

• Statistical matching
  • Ensuring that the treatment and control group are similar pre-treatment
  • Minimizing the effect of unaccounted variables (i.e., confounding variables)
• Counterfactuals
  • Using statistics to estimate the result of being assigned to the opposite group
  • Calculating the true treatment effect
• Multicollinearity, endogeneity
  • Accounting for the possibility of variables and/or the error being correlated with each other
• Hypothesis testing
  • RCTs, ANOVA, T-test, Fisher's test of independence, $\chi^{2}$ test, p-values

Collecting and processing data to build all kinds of machine learning models.

• Classification
  • Logistic regression, KNN, SVM, random forest, gradient boosting, neural networks
• Regression
  • Linear & multiple regression, random forest, lasso & ridge regression, neural networks
• Clustering
  • K-means clustering, fuzzy clustering, hierarchical clustering, density-based clustering
• Dimensionality reduction
  • Principal component analysis, singular value decomposition, linear discriminant analysis
• Deep Learning
  • Feedforward, convolutional & recurrent neural networks, multi-layer perceptron
• Natural Language Processing
  • Tries, named-entity recognition, sentiment analysis, text summarization, topic modelling

Using Bayes Theorem to calculate the conjugate prior, likelihood, and posterior distributions over some hyperparameter.

• Statistical distributions (conjugate prior - likelihood)

  • Beta-Bernoulli / Beta-Binomial distribution
  • Gamma-Poisson distribution
  • Dirichlet-Categorical / Dirichlet-Multinomial distribution
  • Inverse Gamma-Normal distribution
  • Gamma-Exponential distribution

• Generative models

  • Directed graphical models
  • Factor graphs
  • Message passing
  • Sum product algorithm
  • Expectation propagation

• More

  • Using Bayes Theorem to infer probabilities
  • Using Stan to compute samples from a given distribution
  • Selecting appropriate test statistics
  • Calculating confidence intervals

Building computer programs that simulate real-life events and draw conclusions

• Cellular Automata

  • Renormalization, percolation, coarse graining
  • The game of life, the ising model, forest fires, and traffic flow

• Graph Theory

  • Directed and undirected graphs
  • Dijkstra's algorithm, BFS, DFS, random walk
  • Topology / networks
  • Spread of illnesses, information, political and social self-organization

• Algorithms

  • Monte Carlo simulation
  • Markov Chain Monte Carlo (MCMC)
  • Metropolis-Hastings algorithm
  • Gradient Descent
  • Page Rank

Building statistically sound economic models, proving causality, and drawing appropriate conclusions from data

• Causal Inference and Hypothesis Testing

  • Type I and type II errors
  • One tailed and double tailed T-tests
  • ANOVA, Fisher's test of independence, Chi-squared test, p-values
  • Sum of squares total, due to regression, and due to the error (SST, SSR, SSE)

• Fitness of Model

  • Matching (propensity scores, genetic matching, etc)
  • Heteroskedasiticity, confounding variables
  • Multicollinearity, endogeneity

• Econometric Models

  • Regression discontinuity design (RDD)
  • Instrumental variables (IV)
  • Differences in differences (DD)
  • Synthetic controls