Juan D Chacon
A data scientist with a passion for real-world practice. I seek to work in an environment that will challenge me further while allowing me to contribute to the continued growth and success of organizations.
Reach me at: juandados@gmail.com
View My LinkedIn Profile
Work Experience Summary
- Training, Evaluating, Testing & Deploying Data-Driven Models (Scikit-learn, Spark MLlib, TensorFlow)
- Data Mining, Aggregating & Presenting using tools like Pandas, Spark, SQL, & Tableau
- Developing pipelines to ETL data from both real-time and static data sources using PySpark
- Developing dockerized web services using Django, Flask using the TDD paradigm
Education
- Aug 2020 MSc. Applied Mathematics, Simon Fraser University (With Distinction) (GPA 4.18 / 4.33)
- Mar 2016 MSc. Mathematics, Universidad Nacional De Colombia (GPA 4.8 / 5.0)
- Jun 2017 BEng. Electronics, Universidad Distrital FJDC (GPA 4.0 / 5.0)
- Mar 2013 BSc. Mathematics, Universidad Nacional De Colombia (GPA 4.2 / 5.0)
Portfolio
Data Science
Building a predictive model that allows to predict the purchase of 5 banking products, to better guide future campaigns. Challenge description
MeLi Data Challenge [Comming Soon!]
Building a Machine Learning Model to predict the next user purchase based on his historic navigation data. Our approach uses Spark for data ingestion, a text classifier and lots of common sense.
A movie recommendation engine from written reviews using collaborative filtering based on low-rank matrix factorization implemented in Python.
Deep Learning
Gait Inference Using Pose Estimation
A simple methodology for gait prediction based on computer vision that can be used to assist the design process of lower limb active prosthesis. Our proposal uses YOLOv2 for person detection, a human pose estimator build from ResNet-50 and a long-short term memory recurrent network as prediction model.
Other Projects
Safe Coverage of Compact Domains For Second Order Dynamical Systems
A MATLAB implementation of Safe multi-agent coverage for moving domains using artificial potentials for consensus and Hamilton Jacobi collision avoidance.
A Wearable for Gait Estimation
Implementation of a wearable sensor for gait estimation using the MPU 6050 (Accelerometer, Gyroscope, 6 Axis Sensor) on top of an ESP32 (for reading and bluetooth comunication). The data loging and postprocessing is done using python.
