Senior Data Scientist, PhD

Analytics, Machine Learning, Predictive Modeling

Noah specializes in turning complex data into actionable business strategy. His doctoral research focused on optimal decision making with multiple agents of varying performance. He is experienced in developing custom machine learning algorithms for specialized document classification. A combination of Restricted Boltzmann machines, deep learners, and support vector machines feature extraction, dimensionality reduction, and then classification. He has also helped clients by developing multidimensional combinatorial algorithms developed and implemented to exploit market inefficiencies.

Noah's work focuses on five main areas:

1 - Financial modeling: 

  • Algorithmic trading
  • Dynamic pricing and hedging strategies
  • Portfolio optimization
  • Risk analysis

2 - Gaming: 

  • Mathematical models of horse racing , football, cricket, and other games
  • Optimal bet sizing and fund management
  • Arbitrage within and between exchanges

3 - Computational Advertising: 

  • Optimization of advertiser goals using models of consumer behavior
  • Unsupervised demographic identification and targeting
  • Dynamic Pay-Per-Click optimization
  • Complex multi-armed-bandit optimization of content and display

4 - Bitcoins and other cryptographic currency: 

  • Development of BTC based trading algorithms
  • Statistical analysis of blockchain transactions and pricing behavior
  • Author of the Crypto-Currency column for the Wilmott Magazine of Finance

5 - Artificial Intelligence and Machine Learning: 

  • Hierarchical Bayesian models
  • Complex and non-linear inference
  • Natural language processing
  • Document classification
  • Computer Vision 

Selected projects include:

  • Development and implementation of algorithmic trading strategy for commodities futures.
  • Developed improved portfolio allocation strategy for a fund of funds. Utilized MCMC bootstrap of past returns combined with Kelly allocation to find optimal portfolio. 
  • Developed predictive model of student enrollment based on census tract demographics for private educational institution with multiple campuses nationwide.
  • Created recommendation system model for online video based education company.
  • Created commodity price anomaly detection system for commodities trader in Hong Kong.
  • Created system for estimating corporate quarterly earnings as a function of surveyed experts.
  • Converted Excel based boosted-regression stock trading system into R language


Ph.D. UCLA, Statistics

M.S. UCLA Statistics