Machine Learning Engineer | Robotics & Intelligent Systems Engineer

I design and build intelligent systems that connect machine learning, real-time robotics and embedded execution.

My work spans the full AI development stack — from dataset engineering and assisted annotation pipelines to model training, architectural evaluation and integration into robotic systems operating in real environments.

I am particularly focused on perception systems for robotics and applied computer vision.

• End-to-end vision pipeline development (annotation → training → evaluation → deployment)
• YOLO-based detection and segmentation training
• Transformer-based architectures (ViT-MAE, SETR-PUP, SWIN-UNET) experimentation
• Architecture comparison and baseline analysis
• Custom training loops and optimization experimentation
• Model performance evaluation and stress testing
• Dataset engineering and assisted annotation systems

Primary language: Python

• AGV development for indoor and outdoor environments
• ROS and ROS2-based robotic systems
• Navigation stack implementation and experimentation
• Sensor integration and perception pipelines
• Real-time HMI systems for robotic operation
• Firmware development and embedded system integration

Primary language: C++

• Firmware development and electronics integration
• Industrial automation software (B&R systems)
• Embedded AI experimentation on Nvidia Jetson Orin Nano
• Real-time backend systems generated dynamically during execution

Development of a modular framework covering the full machine learning workflow:

• Assisted annotation tools
• YOLO detection and segmentation training
• Transformer-based segmentation experiments
• Multi-model inference
• Dataset stress testing and architectural comparison

• Link: Darkroom-cv

Stack: Python, PyTorch

Full-stack contribution to an autonomous ground vehicle for substation inspection:

• Mechanical and electronics integration
• Firmware development
• Real-time HMI system
• Navigation and control algorithms using ROS

Stack: C++, ROS

Ongoing research involving model optimization strategies and experimentation with training dynamics for improved convergence and performance.

• Vision architecture benchmarking (CNN vs Transformer-based models)
• Dataset quality impact on segmentation performance
• Learning-based perception integration in robotics
• Optimization algorithms for neural network training

Open to roles in:

• Machine Learning Engineering
• Robotics Software Engineering
• Intelligent Systems R&D

• LinkedIn: linkedin.com/in/alecremer
• Email: ale.cremer.dev@gmail.com