Andres Martinez Email and Phone Number

• Led a team of 5 in developing cutting-edge deep learning models using PyTorch and TensorFlow.• Collaborated with team of 20 to carefully hand label over 40,000 images for object detection and facial landmarks, curated the resulting datasets.• Spearheaded a drowsiness monitoring project, integrating object detection, facial landmark detection and pose estimation techniques, achieving over 70% precision in production. Published results and algorithm as the main author in INTERCOM, an IEEE conference.• Evaluated and optimized state-of-the-art lightweight object detection models for deployment on embedded devices.• Upgraded application to the latest version of NVIDIA JetPack (from 4.3 to 4.6), resolving all dependencies and ensuring seamless operation.

• Conducted thorough analysis of various real-time object detection solutions to accurately measure object length.• Proposed the use of forecast algorithms to address problems with latency, poor visibility and low detection confidence in real-time object detection to measure length. Results were presented as a poster in Latinx in AI workshop at ICCV 2021.• Contributed to the development of an aplication that achieved a high level of accuracy in measuring excavator teeth length, consistently within 2 inches of actual measurements.• Developed and implemented a user-friendly image labeling tool utilizing pre-trained models to enhance efficiency in data annotation tasks.

Teaching assistant for physics 1, 2, and 3.