Jose Alarcon

AI/ML, Computer Vision, Data Science

About Me

👋 I turn research into real-world solutions. With a PhD in engineering, and a background in software development and computer vision, my research experience enables me to ramp up quickly, understand technical challenges deeply, and translate cutting-edge ideas into practical applications.


Recently I'm focused on building novel machine learning models - ensuring they will be effective in real-world applications without compromising on safety. To do this I've applied deep learning frameworks like PyTorch and OpenCV in real-time systems, solving challenges with image classification.


I thrive at the intersection of research and engineering, ensuring advanced technologies work outside the lab. Whether it's optimizing a complex data processing workflows on-hardware, training a novel AI/ML model, or working out tracking with 6D vector math, I'm focused on making a tangible impact. If you need someone who can apply AI/ML in complex software and deliver results, let's connect.

Experience

NDI

Northern Digital Inc.

⇾

Research Scientist

  • Led research and development of real-time software systems leveraging machine learning, sensor fusion, and advanced tracking algorithms.
  • Mentored teams on integrating cutting-edge ML and deep learning frameworks (PyTorch, CUDA) into software for clinical workflows.
  • Collaborated with interdisciplinary teams, bridging the gap between research and production.
  • Delivered high-performance, low-latency software optimized for medical imaging and AR/VR applications, ensuring reliability and performance in production environments.
ML/AI Leadership Research Software Metrology Photogrammetry
MIS Electronics

MIS Electronics

⇾

Sensor Specialist

  • Development of an embedded vehicle camera system for 360° "bird's eye" view.
  • Coordinated with teams to integrate hardware and software smoothly.
  • Software development of the image processing needed to perform image stitching.
  • Selected camera sensors and lenses to suit the real-world requirements and constraints.
Computer Vision Camera Calibration Image Processing Optimization Automated Inspection Python OpenCV MATLAB
Vista

Vista Solutions

⇾

Intern, Research & Development

  • Implementation of high-speed laser-scanning system, ensuring high reliability in a fast-paced manufacturing environment.
  • Communicated effectively across teams to align inspection requirements.
  • Took ownership of optimizing OCR and 3D vision pipelines, improving accuracy and efficiency in high-speed inspection.
Computer Vision Machine Vision 3D Image Processing Optical Character Recognition (OCR) Image Recognition Image Acquisition Automated Quality Control Real-Time Processing Algorithm Development Character Segmentation Pattern Recognition C# Halcon
uwindsor

University of Windsor

⇾

Research Assistant

  • Interactive tool for simulating multi-camera networks.
  • SDK to handle manipulation of polygon meshes using the Render Dynamic data structure.
  • Teaching assistant.
  • Multiple paper publications.
Computer Vision Machine Vision 3D Image Processing Image Recognition Image Acquisition Algorithm Development Python Research

Projects

NDI - AI/ML Model For a Medical Workflow

  • Developed a convolutional neural network (CNN) for domain-specific image classification, improving model accuracy by 30% over classical computer vision techniques.
  • Led a successful Proof of Concept (POC) development through comprehensive data collection, model training, and validation, securing stakeholder buy-in with clearly presented results and analysis.
  • Conducted targeted hyperparameter tuning informed by experimental validation and recent machine learning research.
  • Optimized image preprocessing, data augmentation, and feature normalization, enhancing training efficiency, model convergence, and generalization.
  • Developed, trained, and evaluated deep learning models using PyTorch and Python, incorporating design considerations for efficient edge deployment.
  • Improved dataset information density and streamlined data pipelines, significantly boosting model performance and reducing computational overhead.
  • Implemented model evaluation metrics (F1-score, Precision, Recall) for performance benchmarking.
Machine Learning (ML) Deep Learning (DL) Artificial Intelligence Convolutional Neural Networks (CNNs) Image Classification Computer Vision Neural Networks Data Engineering Hyperparameter Tuning Data Augmentation Model Evaluation AI Model Optimization Real-Time AI Processing MLOps Data Pipelines Supervised Learning Research PyTorch Python Tensorflow TorchVision OpenCV Scikit-Learn SciPy Numpy Pandas

NDI - Magnetic Tracker for AR/VR Application

  • Invented and patented (US10539644B1) a method to resolve PLL ambiguity in multi-receiver electromagnetic tracking (EMT) systems using IMU correlation to ensure correct signal demodulation and interpretation.
  • Owned the entire development process, from experimentation to implementation, including algorithm design, firmware integration, system calibration, and production deployment.
  • Engineered real-time signal processing algorithms for a multi-receiver, multi-transmitter system, optimizing for tracking accuracy, and meeting strict processing-time constraints.
  • Built a synthetic data simulator and real-time playback utility, enabling faster debugging, validation, and system refinement for the team.
  • Designed a C++ API for the magnetic 6D tracker, integrating it into virtual and augmented reality applications.
Signal Processing Phase-Locked Loop (PLL) Electromagnetic Tracking (EMT) IMU Correlation Inertial Measurement Acoustics Algorithm Development Multi-Sensor Systems Real-Time Processing Embedded Systems Simulation Experimentation Navigation C++ MATLAB

Vista Solutions - Braille OCR using 3D Laser Scanning

  • Designed a real-time computer vision system for automated pharmaceutical label inspection, delivering a production-grade solution on-site.
  • Delivered a 3D image processing pipeline utilizing a high-accuracy laser scanner for Braille verification at speeds exceeding 100 meters per minute.
  • Engineered an OCR-based label verification algorithm, ensuring that Braille was accurate and readable at all times.
  • Optimized image acquisition and processing, enabling seamless real-time data integration with manufacturing workflows.
Computer Vision Machine Vision 3D Image Processing Optical Character Recognition (OCR) Image Recognition Image Acquisition Automated Quality Control Real-Time Processing Algorithm Development Character Segmentation Pattern Recognition C# Halcon

Private Ventrue - Object Detection for Medication Dispensing

  • Developed a computer vision and object recognition system for automated pill counting, significantly reducing prescription counting errors and manual labor.
  • Built and optimized a real-time data processing pipeline to enhance pill detection accuracy and classification.
  • Delivered a functional Proof-of-Concept (POC), integrating hardware testing rigs and client validation experiments.
Computer Vision Object Detection Machine Vision Automated Inspection Image Recognition Real-Time Image Processing Object Detection Data Processing Algorithm Development Automation Python OpenCV

NDI - Sensor Fusion for Motion Tracking

  • Engineered a robust sensor fusion pipeline combining IMU data with optical tracking, significantly improving system sampling rate and reducing latency.
  • Implemented an extended Kalman filter to estimate motion dynamics, using the IMU to make predictions, and the optical tracker to correct the state.
  • Optimized real-time processing pipelines, improving sampling rates and tracking robustness.
Sensor Fusion IMU Optical Tracking Kalman Filter Motion Prediction Low-Latency Processing Real-Time Processing Time Series Forecasting Predictive Analytics Data Science C++

NDI - Uncertainty Modeling and Compensation for Optical Tracking

  • Implemented a model to quantify and compensate for systematic errors in optical tracking technology.
  • Used nonlinear optimization methods to best fit the model's parameters, improving tracking accuracy under real-world conditions.
  • Integrated uncertainty estimation into production software, enhancing system reliability for manufacturing and medical applications.
Optical Tracking Uncertainty Modeling Nonlinear Optimization Levenberg-Marquardt Calibration Production Integration Data Science Algorithm Development Python Vector Mathematics Computational Geometry Real-Time Processing High-Accuracy Measurement Systems C++ OpenCV

NDI - 3D Surface Reconstruction for a Clinical Workflow

  • Developed 3D reconstruction algorithms integrating photogrammetry and multi-sensor data, enhancing model fidelity and accuracy.
  • Optimized algorithms by integrating CUDA-accelerated computations, increasing frame rate three fold.
  • Implemented a multi-threaded, publisher-subscriber data acquisition system to improve real-time data processing efficiency.
  • Managed CPU-GPU memory allocation and optimized data transfers, reducing performance bottlenecks and improving solution scalability.
Computer Vision CUDA Parallel Computing 3D Reconstruction Photogrammetry Multi-Threading Memory Management Publisher-Subscriber Model Sensor Integration High-Performance Computing (HPC) Data Engineering C++ OpenCV

JobFunnel

  • Build GH actions to deploy Jobfunnel, a job searching tool with 2k+ stars and 300+ forks at the time of writing.
Python Web Scraping Automation Open Source Data Parsing BeautifulSoup Selenium Requests Data Filtering

Education

uwindsor

University of Windsor

⇾

PhD in Electrical and Computer Engineering.

Publications

- Tracking an object in an electromagnetic field

Patent No. US 10,539,644 B1
Jose Alarcon Herrera, Andrew Wiles, Asli Dinc, and Martin Kopcik

- Semi-Automatic Model-Based View Planning for Active Triangulation 3D Inspection Systems

IEEE/ASME Transactions on Mechatronics, vol. 20, no. 2, pp. 799 - 811
A. Mavrinac, X. Chen, and Jose Alarcon Herrera

- 3D Model-based Multi-Camera Deployment: A Recursive Convex Optimization

IEEE/ASME Transactions on Mechatronics, vol. 20, no. 6, pp. 3157 - 3169
X. Zhang, X. Chen, Jose Alarcon Herrera, and Y. Fang

- Viewpoint Selection for Vision Systems in Industrial Inspection

IEEE International Conference on Robotics and Automation, 2014, pp. 4934-4939
Jose Alarcon Herrera, X. Chen, and X. Zhang

Contact