I am a systems and computing engineering from UNAL. I am a passionate software developer with strong computer science foundations. Currently I am working on Cortana, Microsoft's AI powered digital assistant. As a software engineer I develop scalable cloud services and frameworks that power Cortana on different platforms and devices.
Besides software engineering, I am interested in artificial intelligence research both in theory and practice using the latest advances such as deep learning.
Finally, I have experience in real and combinatorial optimization using evolutionary algorithms.
Andrés Felipe Cruz
Android application to automatically detect food from an image and retrieve its nutritional information. This app was developed along with two partners to participate in Tu App (2016).
The image detection component was developed using InceptionV3 (by google), a deep convolutional neural network. We applied transfer learning using the pretrained model in tensorflow. Android, tensorflow, Image detection
Full stack developer in a web application about cognitive stimulation of older adults through memory exercises.
The app was developed using the grails framework and CSS/JS.Web Applications, grails
My undergraduate thesis is an evolutionary algorithm that self adapts its operators at the same time it evolves the solutions. This approach is tested using real benchmark functions and in the degree constrained minimum spanning tree, giving competitive results and interesting insights in the self adaptive process.Optimization, evolutionary computing, genetic programming
I developed a tool to analyze text data recollected using the public APIs from facebook and twitter. The tool contains a small search engine, and a visualization module to show the clusters of topics in the data and some statistics.Data mining, crawling, unsupervised learning
Android application to share music in meetings. Allows collaborative and real time song lists.Android, Mobile development
For our Machine Learning class project me and a friend implemented an agent that learned to play mario kart super circuit (gameboy advance) using deep reinforcement learning. The results were impressive, the agent learn from the raw pixels (only downsampling and gray scale filter is applied) and the reward is the current velocity.Deep Learning, Artificial intelligence
Bachelor of systems and computing engineering • April 2017
My bachelor thesis is about a self-adaptive evolutionary algorithm that evolves the genetic operators at the same time the candidate solutions are optimized. This approach was used to solve the degree constrained minimum spanning tree problem with competitive results.
Natural language processing and text mining • August 2016
Design and analysis of algorithms Part II • 2014
Intro to algorithms, social network analysis • 2014
Systems technician • August 2011
Angela J. Pereira-Morales, Andrés F. Cruz Salinas, Jairo Aponte Melo & Francisco Pereira Manrique• International Journal of Neuroscience · March 2017Link
I have experience using the following technologies in academic and professional proyects.
Any research or software development project? email me at email@example.com