WILBERT PUMACAY
[Link]. Student in Computer Science
[Link]@[Link]
[Link]
EDUCATION
San Pablo Catholic University ( Perú ) , [Link]. 2017 - present
Advisor: PhD José Eduardo Ochoa Luna
Department of Computer Science
National University of Engineering, UNI ( Perú ), B
.Eng. 2008 - 2012
Mechatronics Engineering
Department of Mechanical Engineering
WORK EXPERIENCE
National University of Engineering (UNI), Research Assistant 2012-2013
● Worked implementing C/C++ Hardware Abstraction Layers for microcontrollers and
DSPs, exposing their peripherals through easy to use APIs for various projects.
● Developed required software libraries (in C/C++, MATLAB and LabView) to interact
with actuators (motor-drivers) and sensors (IMUs, LIDAR), used in a differential drive
robotics platform built from scratch (as part of a funded research project).
● Implemented Localization (particle filter based localization using LIDAR), Planning
(A* based path-finding), and Control (sliding-mode and fuzzy) algorithms for the
aforementioned robotics platform (as a personal side project).
Bamtang Games, Software Engineer 2013 - 2016
● Worked as a Gameplay, UI and Engine programmer for the web division of the
studio, writing the required game engines for new games and required gameplay
features (mostly in Javascript).
● Developed tools for artists and designers to integrate their assets into the engine
and speed up their workflow (mostly in C++ and Python).
● Worked as a Gameplay and UI programmer for the console division of the studio,
helping with gameplay features for PS4|XBOX|PC games, and tweaking engine
features for each platform.
SKILLS
C/C++: G
ood working knowledge. Can architect and maintain Projects/APIs (some
projects are listed below)
● OpenGL-based Rendering engine1
● Engine-agnostic simulations2 (under development)
Python: Good working knowledge. Can create packages, add features to codebases and
maintain them (some projects are listed below)
● DeepRL projects from Udacity DeepRL nanodegree3
● Robotics projects from Udacity RoboND nanodegree (navigation4 | kinematics5 |
perception6 | d eep-learning7).
1
[Link]
2
[Link]
3
[Link]
4
[Link]
5
[Link]
6
[Link]
7
[Link]
Javascript: Good working knowledge. Can create web-demos and tools (some projects
are listed below) (games can’t be listed due to privacy policy of the studio’s clients)
● WebGL-based visualizer8, written in Typescript.
● Denavit-Hartenberg playground9, and FK Kuka demo10 (might take a while to
load)
Pytorch|Tensorflow: Working knowledge. Can create models, data-utils, debug
models, present results and debug|fix existing projects (some projects are listed below)
● DeepRL projects from Udacity DeepRL nanodegree
ROS: W orking knowledge. Can design experiments, create tools and use comfortably
the piping provided by ROS. (some projects are listed below)
● Robotics projects from Udacity RoboND nanodegree (kinematics | perception)
● Helped in Udacity’s forum to debug fellow students’ projects.
Embedded systems: W orking knowledge (a bit rusty). Can create HAL libraries for
microcontrollers and DSPs. Comfortable developing libraries from scratch using the
datasheet of a given module|driver|sensor. (some old projects are listed below)
● HAL libraries11 for Texas Instrument’s T M4C123GH6PM ARM-MCU
● HAL libraries for Texas Instrument’s M
12
SP430F5529 MCU
Other: I’ve also worked with the following technologies|libraries:
● OpenCL and CUDA
● OpenCV
● Qt|PyQt|PySide
● Soldering and electronics prototyping skills.
● CAD design (Autodesk Inventor)
PRESENTATIONS
● Poster presentation at [Link]. Presented current progress of my Msc. thesis:
“Loco: A DeepRL framework for Robot Locomotion in Complex Environments”.
● A tutorial on DeepRL at “ I Simposio Peruano de Deep Learning”. Slides h
ere.
MISCELLANEOUS
● Finished Udacity Deep Reinforcement Learning Nanodegree.
● Finished Udacity Robotics Nanodegree (term-1).
● Ap ull-request13 (not merged) for a visualizer14 for Deepmind’s dm_control package.
● A pull-request (not merged) for extra tools15 for Udacity’s Quadrotor Simulator
(made in Unity).
8
[Link]
9
[Link]
10
[Link]
11
[Link]
12
[Link]
13
[Link]
14
[Link]
15
[Link]