Physicist and passionate Data scientist with several years of study and practical experience creating data-driven results for clients. More than seven years of work experience developing projects, applying artificial int…
Physicist and passionate Data scientist with several years of study and practical experience creating data-driven results for clients. More than seven years of work experience developing projects, applying artificial intelligence technologies, machine learning and data science. Experience in design and implementation of analytical models in sectors such as credit risk, mass consumption, pharmacy, transport, education, electricity demand forecast and acoustic models, language modets, sentiment analysis, text mining and development of audio recognition systems and chatbots. Experience in leadership of multidisciplinary teams and under agile methodologies for analytical projects such as CROPS-DM and SCRUM. Experience in the development of applications such as Power BI, Shiny, Dash, SAP Analytics Cloud and Azure and in the model deployment and monitoring with Arise AI and Azure. A responsible person, with high tevel of commitment and persistent. Efficient, with analytical and critical vision, and passionate about learning and innovation.
design models about credit risk, optimizations, fraud, electricity demand. Mlops researcher.bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb
Implemented an optimization algorithm (Simulated annealing) to minimize the CPR indicator.
Met restrictions such as budget, allowed ads, number of ads per week, Trp's, and audience.
Worked the algorithm in Python.
Printed the ads in an Excel box from Python.
Investigated practices and tools to guarantee the correct implementation of MLOps in the company.
Performed testing and deployments with some data sources and company models.
Designed and implemented an S3 classification model for physical and legal clients of the BCC - Banco Cooperativo Cajamar.
Used artificial intelligence techniques and explainability techniques.
Used libraries for manipulating large volumes of data such as Ffbase.
Improved both precision and recall compared to previous models.
Participated in the KDD cup project.
Built a recommender system based solely on user interactions with an ongoing session.
Made predictions for the next element with an acceptable score.
Placed 9th in the competition.
Implemented a model that prints the risk level of an affiliated company leaving the Colsubsidio compensation fund.
Implemented a classification model considering the phases to reach the final state "abandonment".
Achieved a satisfactory f1 score.
Correctly predicted 90% of the companies with the highest contributions categorized as "Gold".
Built an extrapolation engine based on surveys of some affiliates.
Used the universe of affiliates to the box to know their interests and guide advertising and marketing correctly.
Implemented a stratified sample classification and extrapolation model.
Achieved an 85% success rate in guided advertising.
Built a bank credit card fraud model.
Implemented isolation forest to detect anomalies.
Detected 82% of frauds correctly.
Deployed the model in Azure Machine Learning.
Deployed the model to be queried by operation.
Built an audio recognition system that allows qualification of calls between advisors and clients.
Created language and acoustic models from scratch to transcribe audio to text.
Worked on sentiment analysis to rate calls.
Won the championship challenger faced with auditing.
Built a classification model that predicts payment probability to focus efforts on collections.
Built a model for all phases until reaching the state of "agreement" in collections.
Achieved accuracy of more than 85%.
Built a credit granting model for clients of Bancolombia, BBVA, Davivienda and Éxito.
Implemented models with logistic regression and machine learning with explainability techniques.
Guaranteed a healthy portfolio compared to previous methodologies.
Implemented a product recommendation system based on sessions and based on user ratings to increase sales on the web.
Increased sales by 30%.
Implemented statistical models in R and Python including linear regression, decision trees, and random forest.
Applied deep learning techniques to analyze data from FAC projects.
Implemented the INVERSE3D program in the analysis of radiation dose data for long geodetic and polar flights.
Implemented GEANT4 to simulate particle detectors.
Built a time series model to predict the sales of some products in the coming months.
Built an analysis to understand the level of acquisition for certain types of clients.
Worked on classification and segmentation models.
Deployed models in the Azure cloud to be consulted by operation.
Perceived a significant increase in sales.
Reached new potential customers.
Loaded databases with ETL processes to PostgreSQL warehouse.
Updated and maintained availability of certain data.
Gave some roles access to data sources.
Developed a chatbot so that parents on school routes have access to the exact location of their children and information about payments and schedules.
Reduced response time to parents.
Made the route more efficient.
Implemented a machine learning algorithm to identify the areas with the highest traffic at certain times and under certain conditions.
Analyzed geospatial data and created visualizations with H3.
Reduced route travel times.
The AI-first Recruitment Operating System that unifies and automates the entire hiring lifecycle for modern enterprises worldwide.
Dr Job FZ LLC. 2026 © All Rights Reserved
