Mateo Restrepo Mejia

Until Jan 2024, I worked at La Haus, a tech-based residential property market-place in Colombia. Together with my team, we carried out several machine learning / AI related initiatives: such as, a GPT chatbot agent using ChatGPT + RAG, automatic image quality scoring, for ranking of property images in search results, a personalized recommendation engine, to help users find best house for them according to their search parameters and an automated valuation model. I worked for Bancolombia for three and half years as part of the new "Analytic Capabilities Team" (Dirección de Capacidades Analíticas), in the adoption of machine learning based solutions for several of the bank's most pressing problems. Together, we develop analytic models, both supervised and unsupervised, that allow us to extract actionable insights from our data Our group is also closely related to the technology team which supports the Bancolombia's big data lake which runs on Cloudera Distribution for Hadoop (CDH). Before joining Bancolombia, I used to work in the US as "Quant", i.e. a quantitative researcher and developer. During that time I gained 4 years of experience in development of high frequency pricing, risk monitoring, and real-time, deflection-based inventory management for HF trading (listed options market making). I combine expert analytical and mathematical knowledge with effective quantitative application programming skills in a variety of languages. PhD thesis research in simulation-optimization of emergency vehicle systems; real-time redeployment of ambulances. Specialties: Quantitative development in Scala / Python / Java, Quantitative research Matlab / R, SQL Analysis of large data sets, Simulation-optimization, Continuous and Discrete optimization, Mathematical Modeling. Series 7 & Series 63 Certifications. Probability. Symbolic computation with Schur functions.