Stuart Lempel

Software Engineer, Data Scientist, and motivating leader with the ability to guide client ideas from inception to goals and deliverables. Holds an active top secret clearance, with mastery understanding of designing and implementing database schemas for optimal system performance and efficient post processing extraction. Used a wide variety of development tools, and over the last 10 years have primarily focused on Ruby on Rails, Python, and Java. Worked on many long-term projects, and also built many application demos and proofs of concept.

Top Secret

Tennis Handicap

Is it possible to use historical data and AI algorithms to gain statistical advantage in predicting the winner of each tennis serve?

• Built a dataset of every available tennis serve since for the last 20+ years.
• Built an automated system that followed real-time every current major tennis match.
• Was able to gain advantage over odds of each point with the AI and the historical data.

Machine Learning Ruby Gem

Ported many of python's Scikit-Learn packages to ruby

• Decision tree: cart, random forest
• Cluster: k means, k nearest neighbors
• Regression: linear, logistic

Lending Club

Is it possible to predict loan repayments or defaults based off of other criteria than credit report and credit history?

• Built and trained a decision tree model based off of key words in the reason for the loan.
• Result showed a hirer rate of late payments or defaults for toys such as jet-ski or motorcycle.
• Also fond a higher rate of loan repayments for family things like nursery or daycare.

Stock Sentiment

Can overall change in message sentiment of a stock predict buy and sell points?

• Scraped websites to build a sentiment database of stocks per day.
• Trained the data sentiment and price change per day.
• Was able to predict buy and sell days based on current sentiment.