3 years 8 months in Wolfe's QES team. I built factor, ML and NLP signals delivered to institutional clients, including all of QES's network-momentum signals. I led 2 QES reports, contributed to 7, and presented the research to clients and PMs.
Researched intangible assets as a risk factor that book-value factors miss. Backtests: loss-making stocks model, Sharpe 1.53; NLP long-short signals, Sharpe above 1.
I combined fundamental, technical, and alternative signals in a risk-neutralized factor model using an ensemble of XGBoost and Elastic Net. I analyzed what drove factor performance and used an event study to measure how markets reacted to corporate turnarounds.
R&D and SG&A are expensed, so book value, and with it book-to-price, misses intangible capital, most of all in tech and health care. We capitalized both into an intangible-intensity measure and tested it as an added style factor in the QES risk model.
I built an NLP system that analyzed earnings-call transcripts and audio, 10-K and 10-Q filings, and news. I used transformer models including RoBERTa, Llama, and GPT, alongside knowledge graphs, to score sentiment, model topics, and recognize entities.
I built cross-sectional exposure screens and scenario analyses to show which equities and portfolios were sensitive to shocks involving interest rates, oil, geopolitics, or technology.
I led the design of dashboards for researchers to run backtests, screen stocks dynamically, and evaluate trading strategies.