Data Science | Machine Learning, Risk Modelling & Econometrics | BSc Economics

Economist turned data scientist, working where econometrics meets machine learning: predictive modelling, inference and segmentation on tabular data. I build models to support a decision, and document the reasoning as carefully as the result.

📍 Based in Santander, Spain
🏢 Data Science Intern at Laborare SLP (legaltech startup)
🎓 BSc Economics, Universidad de Cantabria
🌍 Erasmus+ at Vrije Universiteit Brussel (Solvay)
📈 MSc in Data Science (UC-UIMP), 2026-2027 cohort, coordinated by IFCA (Institute of Physics of Cantabria), a centre of the CSIC (Spanish National Research Council)

Machine Learning & Modelling

Data & Analysis

Visualisation & Delivery

Tools

• credit-risk-modeling, an end-to-end expected loss pipeline (PD × LGD × EAD) on a real loan portfolio. 95% recall on defaulters, 4.49% estimated portfolio loss, deployed as a Streamlit scoring app.
• yield-curve-forecasting, Nelson-Siegel curve fitting on FRED and ECB data, with VAR/LSTM forecasts and a recession classifier validated through walk-forward backtesting. (collaborative, in progress)

• gdp-convergence-analysis, cross-country test of β-convergence in real GDP per capita across 153 economies (World Bank, 2004–2024). A result of β = −0.354 (p < 0.001) implies a convergence speed of just 0.37%/yr, an order of magnitude below the canonical 2% benchmark. σ-convergence confirms the distribution barely narrows. Conditional and convergence-club specifications, HC1 errors and influence diagnostics.

• bank-campaign-predictor, a term-deposit subscription model tuned for recall over raw accuracy: AUC 0.921, 90.7% recall on subscribers.
• income-segmentation-r, PCA + K-Means on World Bank indicators for 169 economies. Three components explain 88.67% of variance and recover the official income classification, validated with Silhouette and Davies-Bouldin.

Bachelor's Thesis, "Technological Innovation in Spanish Manufacturing" (R). Balanced panel of 1,355 firms and 10,840 observations from PITEC (2008–2016). Fixed and random effects under one-way and two-way specifications, Hausman tests for estimator selection, lagged regressors and two dependent variables as a robustness check. Grade: 8.9/10.

Data Science Intern | Laborare SLP | June 2026 - present

• Analysed +100 survey responses by building an end-to-end Python pipeline: cleaning, Welch t-tests, Wilson intervals, willingness-to-pay curves
• Deployed the survey landing page to production as the MVP's data-collection layer
• Instrumented the validation funnel: 36 KPIs with formulas, sources and owners, UTM schema, survey variable design

BSc Economics | Universidad de Cantabria | 2021 - 2026

• Relevant Classes: Econometrics, Statistics, Microeconomics, Macroeconomics, Multivariate Data Analysis

Erasmus+ Exchange | Vrije Universiteit Brussel, Solvay Business School | 2024 - 2025

• Relevant Classes: IT Modelling, Programming, European Economics, Operations Management

Cantabria Tech Talent | Data Analysis programme (UNIR) | 2026

• Awarded one of 68 places out of 768 applicants

Certifications

• Deep Learning Specialization, DeepLearning.AI
• Machine Learning Specialization, DeepLearning.AI
• Financial Engineering and Risk Management, Coursera (by Columbia University)
• Business Analytics with Excel, Coursera (by Johns Hopkins University)

• Machine Learning
• Credit and model risk
• Time series forecasting (ARIMA, VAR, LSTM)
• Causal inference
• Econometrics

• 📧 Email: sergioiglesiaslopez03@gmail.com
• 💼 LinkedIn: linkedin.com/in/sergioiglesiaslopez
• 🐙 GitHub: @sergioiglesias1

_{All projects are built from scratch and fully documented.}