Jonathan Mcwilliams Email & Phone Number

Mountain View, Ca, Us

San Francisco, California, Us

Bio: https://www.kaggle.com/about/team

• Led implementation of Google’s Tensorflow Wide & Deep neural net recommender which was integrated with Databricks Feature Store for uniform feature development coded in Python• Utilized structured streaming to aggregate silver tables to a single gold on Databricks on AWS for use in production-level real-time data streaming coded in Spark SQL• Led development of AWS Transcribe and AWS QuickSight in Python to conform with our existing data structure

• Provided liquidity for protocols through creation and staking of LPs on most CEXs, L1 chains, and major DEXs• Utilized Web3.py in Python to automate reading and writing to several blockchains• Programmatically swapped L2 tokens into L1 tokens and reinvesting funds into yield bearing LPs coded in Python• Analyzed Masterchef contracts to verify nonexistence of malicious code within fork coded in Python

Bellevue, Wa, Us

• Led development from concept to production of PacBot cloud compliance model which classified cloud instances’ security threats• Developed within Databricks on AWS in Python and Spark SQL. Shared trained model with developers using joblib dump of sklearn fit pipeline for deployment of model coded in Python• Taught a weekly Introduction to Machine Learning class to junior software engineers so they had a high-level understanding of ML and its use cases

San Jose, Ca, Us

• Revamped forecasting process within marketing department using Facebook’s FBProphet time series forecasting model coded in Python• Built dashboard which highlighted significant WoW movements on tracked metrics aggregated coded in SQL• Managed the Weekly Business Report (WBR) outlining prior week’s performance; led weekly meeting discussing key findings with executive stakeholders• Managed monthly outlook process which tracked QTD performance vs forecast, specifically bringing attention to regional over/underspend vs forecast

Redmond, Washington, Us

• Built solutions on Databricks in Azure in Python, transferred data to and from Azure blob storage containers coded in Spark SQL • Wrote various output to MongoDB in JSON format for API integration• Developed classification model which identified contracts requiring human review• Utilized several NLP tools to identify problematic words or phrases coded in Python• Wrote and modified several SQL queries within Microsoft’s SSMS platform• Reduced 9,000 FTE hours previously allocated towards manually reviewing contracts• Conducted code reviews and led several offshore-onshore syncs

Seattle, Wa, Us

• Developed XGBoost model for a classification model coded in Python which identified phone calls requiring human review• Utilized IBM Watson Speech-to-Text API to transcribe 17,500 calls / 600+ hours of customer service calls coded in Python• Utilized several NLP tools in Python to gather sentiment and build a classifier system to determine which calls lead to positive or negative customer service experiences• Linked call surveys, previously graded by a human, as labels for training ML models

Boulder, Colorado, Us

• Capstone Project: Built a classifier which predicted whether companies in the S&P500 would will beat their projected quarterly revenues based on the content of their quarterly earnings conference calls. Using numerous NLP tools, including scikit-learn's TF-IDF vectorizer and NLTK’s Vader sentiment analyzer, I created an ensemble of a gradient boosted classifier with a logistic regressor. Project on GitHub: https://goo.gl/x7UNCX• Case Study: Predicted churn in a ridesharing app. Used profit curves, LTV, and a Gradient Boosting Classifier to optimally set thresholds of churn prediction• Case Study: Used text to classify app descriptions as sports related/not sports related. Used NLP via TF-IDF with Multinomial Naive Bayes

Omaha, Ne, Us

• Ranked #1 within entire company in increased assets under management (AUM), $25.7 million, during last quarter of employment• Promoted from Financial Consultant to Senior Financial Consultant after prolonged period of exceeding expectations• Worked directly with clients - oftentimes face to face. Gained understanding of their financial goals and delivered customized solutions• Lead company-wide sales guidance conference call – coached junior Financial Consultants on best practices• Created automated trading algorithms coded in Python• Frequently conducted Monte Carlo simulations on clients’ portfolios to assign confidence intervals of asset longevity• Managed $300 million for roughly 450 clients

Omaha, Ne, Us

Minneapolis, Mn, Us

• Coded in Python and SQL, utilizing BeautifulSoup and Selenium to automate gathering of information by scraping FL.gov layoff reports for 401k rollovers• Raised AUM by $5 million within first 12 months of hire

New York, Ny, Us

• Managed $200 million for 100 affluent households • Recognized in the Big Bull Rankings - the daily top 5 highest grossing advisors in southern Florida• Awarded the Silver Bull for opening a $250,000+ relationship within first three months of hire

St. Louis, Mo, Us

• Designated as a Segment Leader in region for having highest gross production in segment• Raised assets under management by $3 million within first 12 months of hire

Certificate, Data Science

Bachelor Of Science, Finance