![LG](https://d2abrqas732254.cloudfront.net/images/specialists/ccbfa6b3bb25974d5b500bffbb85a59f232ea336.png)
English:
Upper Intermediate
Luis G.
Vetted by YouTeam
Uruguay
UTC -03:00
America/Montevideo
English:
Upper Intermediate
Seasoned ML & CV Engineer / Technical Lead
Overall 10 years of work experience. Ph.D. in Neural Networks. Expertise in AI and soft computing approaches (Artificial neural networks: CNN, Yolo_v3, MobileNet, Dlib models and custom models, genetic algorithms and machine learning: k-means, k-nn, SVM and so on). Expertise in C/C++11, Computer vision (HOG, OpticFlow, Homography), OpenCV, automotive, ADAS, Embedded Systems, OpenGL (including shader language GLSL), GPU drivers development, Linux kernel drivers, Audio Video pipeline (codecs, containers). Wide experience in system analysis, troubleshooting and debugging.
Want to hire this engineer?
Check if Luis is availableExpertise
Years of commercial development experience
10 years of experience
Core technologies
Other technologies
Project Highlights
Face recognition in retail
The project aims to develop the system which automatically measures customer's satisfaction level, age/gender distribution and warehouse spots of the interest.
Responsibilities & achievements
- Client architecture creation - Full client CV pipeline construction - Technology strategy and client data pipeline definition for: * face detection * face/person tracking * face recognition * emotion recognition * age recognition * gender recognition * sight direction determination * statistic aggregation - Performance optimization
Parcel station
Algorithm for automating the process of recognizing packages using neural networks. The program attempts to recognize the postal label on an image obtained from the user, searches for the label in the database and sends all necessary information to the user.
Responsibilities & achievements
Increased system recognition accuracy Decreased false positive detection rate
Connected Car
The project is intended for measuring the driver's behavior and scoring driver for his rides. The project consists of several parts - a hardware device that allows measuring driver's behavior on 4 parameters - hard braking, hard acceleration, hard cornering, and speed. The other two parts are WebView that serves as the admin panel for data management and mobile application where the driver can see his statistics and scoring.
Responsibilities & achievements
Client architecture design Initial IoT device evaluation Sensor signal filtration Sensor signal processing Fourier transformation for sensor signal Signal pattern classification
Under NDA
Implementation and improvement of various computer vision techniques for tracking and analyzing players and game ball trajectory and statistics for the video streaming platforms. The ball recognition using Homography, OpticFlow and Yolo v3. The game field landmarks recognition. The UFC fighters recognition by pose estimation. The face landmarks detection and face emotion recognition using Dlib and specially trained CNN. The solutions were using various OpenCV methods, Darknet, Tensorflow.
Responsibilities & achievements
Improvement for the object rectangle refinement, custom approach for trajectory analysis, performance analysis and improvement Development of the custom neural network training framework using Tensorflow Face alignment (DLib, OpenCV) and so on Code development on Python and C++14, prototyping
ADAS for R-Car M3/H3
Advanced Driver Assistance System (ADAS), the driver sight direction recognition. The human eye extraction and direction recognition using HOG and SVM models implemented in a special ML HW accelerator. The camera image rectification using advanced HW accelerator. New GPU evaluation, writing benchmarks (GPU and Computer vision-related), quick prototyping for the customer, implementing customer's computer vision-related algorithms on the existent image processing units.
Responsibilities & achievements
Code development on C/C++11 SW System requirements specification/definition (GPU, HW accelerators, Video codecs) SW System architecture review (GPU, HW accelerators) Description of the existent internal SW architecture for GPU and HW accelerators GPU benchmarks development for Mali and Imagination GPUs (from scratch, completely written by me) Code documentation writing CV related algorithms (human eye recognition from camera sensor) benchmarking for R-Car Gen3 (Renesas) IMP IP
Mali Mimir GPU (tMIx), Mali Midgard GPU (t83x, t86x),Vulkan
Investigation of the possible GPU utilization for deep neural networks (CNN) computations. Adding/porting new features, troubleshooting, legacy userspace DDK maintenance and port to the new Mali GPU architecture (Bifrost). Implementation of Vulkan API for existing GPUs (Midgard arch). Porting old DDK onto new SW stack
Responsibilities & achievements
Code development on C/C++1 GPU DDK development for Mali GPUs New GPU features deployment New API implementation and prototyping SW modules porting Khronos CTS verification, troubleshooting, system analysis, bug fixing Troubleshooting in different areas of the SW, including GLSL shader disassembling and GPU debugging Full-stack analysis, starting from the wrong pixel on the screen up to the wrong register/instruction in the GPU or incorrect behavior in the application GPU benchmarking
Sony Xperia Mint, Sony Xperia Z, Sony Xperia Z1, Sony Xperia Z2, Sony Xperia Z3
Fully automated test rig (device, chamber, and software) for the phone frames recognition using a high frame rate camera (120 fps) and OpenCV. Fully automated test rig (device, chamber, and software) for the phone touch latency measurement using a high frame rate camera (120 fps) and OpenCV. Augmented reality application for the camera smartphone, using the Vuforia SDK.
Responsibilities & achievements
Code development on C, C++, and Java Graphics drivers maintenance on Linux Kernel level, Android native level, Android Java level Troubleshooting in different areas: graphics, memory leaks, performance, touch latency, Android OS Crash analysis Touch lag and power consumption New features/concepts/ideas prototyping Taking partial influence in innovations promotion Dev-t Sony applications for efficient OpenGL ES utilization in Android OS Collaboration with a team of 20 members
Education
Higher education in Computer Science
Agency
400+
GMT-11
Remote
Want to hire this engineer?
Check if Luis is available