Severus Snape
Data Scientist · Forecasting & Pricing
severus.snape@example.com · +34 91 555 02 66
Madrid · github.com/ssnape-example
In one paragraph
Data scientist, 8 years, focused on demand forecasting and price optimization for retail. My models set prices on 300k SKUs daily. I care about backtests, holdout honesty, and whether the ops team can actually run the thing at 6 a.m.
Numbers
+3.8%gross margin from pricing engine, yr 1
300kSKUs priced daily by my models
Experience
Lead Data Scientist2022 – present
Mercalia (grocery retail, 1,100 stores)
- Built the markdown-pricing engine (gradient boosting + rules layer); waste down 14%, margin up 3.8% in year one
- Run a team of 4; introduced weekly backtest reviews, model rollback rate fell from monthly to twice a year
- Moved batch scoring from 6h Spark jobs to 40min on DuckDB + orchestration, saving €90k/yr in compute
Data Scientist2019 – 2022
Voltia Energy (utility)
- Short-term load forecasting for 2.3M meters; MAPE improved from 6.1% to 4.2%, feeding daily trading desk decisions
Analyst2017 – 2019
BBVA, risk analytics rotation
- Credit scoring model maintenance and regulatory documentation (IFRS 9)
Toolbox
Pythonscikit-learnLightGBMSQL
DuckDBAirflowdbtSpark
Bayesian methodsCausal inference
Education & talks
M.Sc. Statistics, Universidad Carlos III de Madrid, 2017
B.Sc. Mathematics, Universidad de Sevilla, 2015
Talk: "Backtests That Don't Lie to You," PyData Madrid 2024
Kaggle: 2 silver medals, forecasting competitions