Aug 2022 – Mar 2024 · 1 yr 7 mos

• Hedging optimisation, algorithmic hedging.
• Regression models using lightGBM.
• Optimization using linear programming, gradient based optimisation and differential evolution.
• Detection of states using hidden Markov chains.
• Dash based applications.
• Skills/tools used: Python, Unix, C++, Google Cloud Storage, BigQuery, Spark, tick data

May 2021 – Aug 2022 · 1 yr 3 mos · Warsaw, Mazowieckie, Poland

• Counterparty Credit Risk - production processes for correlations and volatilities for FX, equities, indexes, rates, commodities and inflation. Mainly calibration of OU mean-reverting models.

Nov 2020 – May 2021 · 6 mos

• IB division - Applied NLP for named entity recognition in swap contracts

Mar 2019 – Oct 2020 · 1 yr 7 mos

• Hands-on research on anomaly detection in transactions. Autoencoders, density-based clustering and decision trees. Attribution for neural networks based models (explainable models). Using R, Python (Tensorflow), Hive and GCP.
• At the same time, entity manager of a team that grew from 3 to 18 data scientists.

Oct 2017 – Mar 2019 · 1 yr 5 mos

Jun 2016 – Sep 2017 · 1 yr 3 mos · Santiago Province, Chile

• Internal consultant providing a wide range of analytical solutions to be used by senior management. Some of my tasks were:
• Income modelling for onboarding purposes.
• Portfolio performance analysis.
• Portfolio concentration analysis.
• Analysis of competitor's performance.
• PD and LGD stress test.
• Cost of capital estimation.
• Tools used: R and SQL.

Jul 2015 – Jun 2016 · 11 mos

• Consultant in economics and strategy for clients from Japan, Mexico, US and Chile.
• Antitrust and disputes consulting. Valuations and strategic recommendations. Econometrics.
• As external consultant after finishing the full-time period.

Jul 2014 – Mar 2015 · 8 mos

• C++ code for markov perfect equilibrium model (symmetric equilibrium), the equilibrium was numerically estimated using a reinforcement learning methodology in which agents were learning from their actions and the effects on the environment. Market micro structure model. Work in collaboration with researchers from the UK and France.

Jan 2013 – Mar 2014 · 1 yr 2 mos · Agustinas 975, Santiago

• Quantitative research for models under AMA approach, namely Monte Carlo and Fast Fourier Transform methods for the computation of convolution between severity distribution and frequency distribution of operational risk events.