ÁNGEL CEBALLOS
Querétaro City
Summary
Results-driven and highly skilled Algorithms Engineer with over 3 years of experience in developing and productizing deep learning solutions. Proven expertise in training, fine-tuning, and optimizing perception DNNs in low precision (FP16/INT8) and applying various quantization methods. Adept at improving DNN architectures on NVIDIA GPUs, enhancing inference speed, accuracy, and power consumption. Experienced in leading perception stack projects for ADAS applications, developing PoCs and optimizing computer vision algorithms for autonomous driving functions. Strong background in using deep learning frameworks such as PyTorch and ONNX, with advanced proficiency in Python and C/C++. Seeking to leverage my background in deep learning and software engineering to contribute to the innovative team at NVIDIA.
Overview
4
4
years of professional experience
1
1
Certification
Work History
ADAS R&D Perception Component Owner
Continental
Querétaro City, Querétaro, Mexico
03.2024 - Current
- Leading Algorithms and Testing team, planning and execution of activities
- Research and Develop CNN architectures for target hardware (TI SoC) considering hardware constrains, resource allocations, runtime execution and compatibility with CV stack algorithms.
- Optimize, compile and profile int8 CNN model
- Enhanced perception network while reducing runtime
- Embedded Pipeline: camera intake and processing, perception and filtering algorithms, camera model and post-processing
ADAS R&D CV Algorithms Engineer
Continental
Querétaro City, Querétaro, Mexico
11.2022 - 05.2024
- Data exploration, preprocessing and labeling
- R&D computer vision models: (classification, object detection, semantic segmentation, key-points)
- Improved CNN perception model for trailering function enabling demo function. Key modifications: CNN architecture, data augmentation, labeling-training procedure)
- Model testing, optimization and deployment
- Spatial transformations (2D->3D and 3D->2D)
- Agile planning
- Processing and visualization of 3D radar points
ADAS R&D Algorithm Test Engineer CV
Continental
Querétaro City, Querétaro, Mexico
10.2021 - 11.2022
- Data and sensor validation
- Algorithm testing (perception and trajectory planning)
- Create pipeline for automatic testing and generation of KPIs report
Computer Vision Trainee
GestaLabs
Monterrey, Nuevo León, Mexico
05.2021 - 09.2021
- Data preprocessing and labeling
- Research and compare CV algorithms (OCR, classification, detection, segmentation)
- Research and develop data pipeline for CV solution
- Visualization and interpretation of DNN models
Education
Bachelor of Science - Computer Science
Tecnológico De Monterrey (ITESM)
Monterrey, Nuevo León, Mexico12-2020
Skills
- AI
- Computer Vision
- C, C, Python
- Ubuntu
- PyTorch, ONNX
- Git
- Docker
- Cuda, TensorRT
Projects
- ADAS, 01/01/24, Leading perception stack for parking component in application project for an OEM.
- ADAS, 01/01/24, Developed (PoC) for analyzing radar points/objects caused by multi-path phenomenon.
- ADAS, 01/01/23, Conducted research of BEV architecture and presented key designs plus optimization ideas.
- ADAS, 01/01/23, Developed (PoC) trailering function for autonomous truck based on SOTA key-points pose.
- ADAS, 01/01/22, Improved computer vision algorithm for trailering function.
- ADAS, 01/01/22, Test the algorithms for computer vision and driving function components, evaluate KPIs and automate the process.
- Smart Parking/Security System, 01/01/21, Optimized object detector (Yolov4-tiny) running on Jetson Nano that detects cars and license plate. Consequently extracting the plate information with OCR {EasyOCR, Tesseract, Plate Recognizer}.
- Crop Damage in GeoTiffs, 01/01/21, Detect and quantify crop damage through semantic segmentation.
- VantTec, 01/01/20, Won international autonomous competition Roboboat 2020.
Certification
- Getting Started with DeepStream for Video Analytics on Jetson Nano: https://learn.nvidia.com/certificates?id=49b76c123b024fd788ab829b079dd2b1
- Getting Started with DeepStream for Video Analytics on Jetson Nano: https://learn.nvidia.com/certificates?id=2436819aa3f14a9ab19baf3b42ab513b
- Jetson AI Specialist: https://drive.google.com/file/d/1ihoOu9RZlI6GqO6sGg9aVraoJvB5KAte/view
Languages
English
Websites
- https://www.linkedin.com/posts/continental_autonomousmobility-continentalautomotive-ugcPost-72592292
- https://github.com/Angel-Ceballos/production_challenge
- https://github.com/Angel-Ceballos/smart-parking
- https://docs.google.com/presentation/d/1JwqgqmnkXbTghYEc4TULeECjwIooyNTfaKXYx8xQ_nk/edit#slide=id.p
Hardwareskills
- Lab HW (oscilloscope, multimeter, protoboard, analog-digital IC)
- Nvidia and Texas Instruments Dev Boards
Timeline
ADAS R&D Perception Component Owner
Continental
03.2024 - Current
ADAS R&D CV Algorithms Engineer
Continental
11.2022 - 05.2024
ADAS R&D Algorithm Test Engineer CV
Continental
10.2021 - 11.2022
Computer Vision Trainee
GestaLabs
05.2021 - 09.2021
Bachelor of Science - Computer Science
Tecnológico De Monterrey (ITESM)