BlueVal is pleased to announce that Filippo Leone, Associate at BlueVal Group, has successfully completed his Master’s thesis focused on the application of machine learning and natural language processing techniques to venture capital decision-making.

The thesis was completed as part of a dual-degree Master’s program between the University of Illinois Chicago and Politecnico di Milano, within a research project led by Professor Theja Tulabandhula, and supervised jointly with Professor Rocco Roberto Mosconi.

The research addresses a structural challenge in the valuation of early-stage enterprises: the limited availability and reliability of high-quality quantitative information. At the earliest stages of their lifecycle, startups often lack standardized and audited financial reporting, historical performance data, and other structured indicators typically used in valuation and investment analysis. As a result, traditional quantitative models may fail to fully capture their underlying value, strategic positioning, and growth potential.

To overcome this limitation, the thesis investigates whether the integration of unstructured textual information can complement sparse quantitative data and provide a more faithful representation of early-stage startup conditions. Using a large-scale dataset of U.S. venture-backed companies, the study develops a hybrid predictive framework that combines structured financial variables with semantic embeddings derived from startup descriptions, competitive narratives, and investor-related text.

The empirical results show that incorporating textual variables leads to a measurable improvement in predictive performance when forecasting follow-on funding outcomes, compared to models relying exclusively on structured data. By capturing qualitative signals that reflect vision, market positioning, and perceived credibility, the proposed approach offers a richer and more realistic assessment of startups operating under conditions of high uncertainty and information asymmetry.

This research aligns with BlueVal’s ongoing commitment to innovation in valuation and financial analysis, and highlights the potential of advanced analytical techniques to support decision-making in contexts where traditional data sources are inherently limited.

View Thesis | View Presentation | View Executive Summary