Chicago, Illinois, United States
• Lead Developer for TSecDB Smart Order Router • Quantitatively driven router giving clients access to dark and lit venues, internal liquidity, and market makers, supporting equities, options, and swaps. Optimized for fill rates, price improvement, and fees. • Investigate and delineate aberrant routing behavior in a concise and timely manner for clients; propose and implement solutions to problems • Aggregate and summarize trading performance (TCA) and present to partners and stakeholders • Routinely perform code releases, mantain QA enviroment and 700+ active regtests
• Automated workflow to extract trade data and built customized size and price limits for client block trades, increasing serviceable hours by 60% percent and expanding current client base for electronic trade executions • Collaborated with Mortgage Traders to identify features relevant to client trading patterns for mortgage backed securities (MBS) to construct time series feature matrix for fast data visualization and ad hoc analysis • Trained Machine Learning models to predict trading behavior for 363 clients across 62 types of MBS transactions, increasing sales conversion rate and improving trade strategies
MoSeeker is a Chinese tech firm that serves as a company’s HR service. It uses social media and data analytics to improve the user experience and matching accuracy for both the companies and the job seekers. • Wrote python scripts to extract job applications from SQL server, clean data, and create term frequency and SVD reduced datasets, allowing researchers to seamlessly train predictive models with 400+ features • Developed recommendation system to suggest jobs for users from approximately 200,000 open positions for 500,000 user profiles, using item based collaborative filtering, similarity between profile information and job description, and latent semantic indexing, resulting in enhanced user experience • Built algorithm for filtering submitted resumes for any open job position using field-aware factorization machines on training data of 10,000 historical applications, simplifying the resume selection process for the 500+ client companies
• Served as the statistician for the Competency Model Project that attempts to optimize resource allocation, enhance objectivity in performance evaluation, and provide quantitative basis for bonuses and promotions • Used statistical models to discover and analyze relationships in evaluation data with over 40 explanatory variables across nearly 200 employees, leading to three key departmental changes and proposal to extend evaluation model across other sectors • Developed Excel VBA Macros to automatically generate reports of performance feedback and training progression on over 40 different categories, saving several days of verification, analysis, and documentation per month