MELANY S. ROSALES Woodside, NY ***** 917-***-**** acxfwr@r.postjobfree.com US Citizen

C++ PROGRAMMER COMPUTER VISION ENGINEER

Innovative professional with 6+ years of experience using C++ researching and developing computer vision, image processing, and modeling algorithms. Used STLs when applicable. Worked on other projects such as Android development and Qt GUIs. Can also make webpages. Adept at designing, implementing, and testing applications and providing ongoing support and maintenance for existing software. A team player able to produce results with a group or independently. TECHICAL SKILLS: Languages: C/C++, Java, Bash, CUDA, PHP, JavaScript, SQL, CSS, HTML Software: GIMP OS: Ubuntu CAREER NARRATIVE

PROGRAMMER AT&T (Acquired DirecTV), Stamford, CT 7/2015 to 6/2016

Improved internal prototype Android App. Used Java.

Participated in the logo detection of posters. Used C++ to encapsulate the logo detector. PROGRAMMER DirecTV (Acquired Liveclips/Play It Over), Stamford, CT 3/2014 to 7/2015

Created a face detector utilizing HOG features and Support Vector Machine (SVM). Used C++.

Analyzed and identified 3D angles of a 2D face with specified points on the face. Used C++.

Implemented seam craving. Used tbb (threading building blocks) for multiprocessing in C++.

Wrote two GUIs made using C++ Qt. One was used for annotation of faces and the other for monitoring.

Applied JQuery to create a timeline for incoming football highlights. PROGRAMMER Liveclips/Play It Over, Stamford, CT 11/2009 to 3/2014

Contributed to the development of several C++ computer vision algorithms mostly for the automated system for DirecTV’s Sunday Ticket

Designed a trainable Optical Character Recognition (OCR) that completes identification by gathering features and then feeds features to a random tree classifier to read the mini boards.

Detected scene changes by implementing different methods to determine the most effective to be histograms and edge change ratio.

Filtered out grass pixels and analyzed the histograms and lines on the screen to determine if on the main camera.

Exposed wipes by finding the wipe pattern within the video.

Defined direction of camera motion using dense and sparse optical flow.

Distinguished faces by applying OpenCV's face detect algorithm.

Established location of view in the field (left or right) through lines and histograms analysis.

Tailored algorithms to different sports such as football, soccer, and baseball and targeted different clients which included college football and the NFL. Used C++ classes to categorize the different sports and clients.

Integrated computer vision algorithms into the entire system and managed the consistent change which evolved from VLC to Pronto and eventually Nighthawk.

Applied CUDA to increase algorithm speed to the wipe detector algorithm.

Developed Android App prototype with the capability to receive game alerts through MQTT with an interactive video player. Used Java.

Made a control panel in HTML for the automated system. WEBSITE: www.dbescomics.com

EDUCATION

Machine Learning, Coursera Online Courses, Stanford University 2015 Bachelor of Science, Computer Science, cum laude, University of the Philippines, Quezon City, Philippines 2009 Thesis: Theft Alert of Laptops (TALa), C++ prototype –Links laptop and owner in CCTV footage

Contact this candidate