The posting
We are looking for a Machine Learning Engineer with strong practical experience in computer vision and neural networks to join our team.
The successful candidate will work on the research, development, training, optimization, and deployment of machine learning systems for current and future FluxPose hardware products. This is a hands-on technical role covering the full development cycle, from experimentation and dataset creation to deployment on production hardware.
We are looking for someone capable of independently solving technically uncertain problems and turning research or prototypes into reliable real-time systems.
Responsibilities
- Design, train, evaluate, and improve machine learning models for real-time sensing and tracking applications.
- Develop computer vision and image-processing systems.
- Build and maintain datasets, data collection pipelines, and evaluation tools.
- Generate synthetic training data where appropriate.
- Optimize models for accuracy, latency, memory usage, and computational cost.
- Quantize and deploy neural networks to embedded or hardware-accelerated platforms.
- Develop tools for training, visualization, benchmarking, and automated evaluation.
- Investigate real-world failure cases and improve model robustness.
- Collaborate closely with embedded, electronics, mechanical, and software engineers.
Mandatory Requirements
- Strong practical knowledge of machine learning and neural networks.
- Strong Python skills.
- Experience with PyTorch or a similar modern ML framework.
- Good understanding of computer vision and image processing.
- Experience training, evaluating, and debugging neural networks.
- Good understanding of datasets, augmentation, validation, overfitting, and model generalization.
- Ability to work independently and take ownership of complex technical problems.
- Strong problem-solving skills and attention to detail.
- Fluent English.
Major Advantages
Experience with Edge AI, TinyML, or embedded neural-network deployment is a significant advantage.
Highly valued experience includes:
- INT8 quantization and quantization-aware training.
- Model compression, pruning, or knowledge distillation.
- Compact CNN architectures and real-time inference.
- ONNX, TensorFlow Lite, TensorFlow Lite Micro, or similar technologies.
- NPUs, DSPs, GPUs, or neural-network accelerators.
- C# experience.
- C or C++ development.
- Microcontrollers and embedded systems.
- Real-time computer vision.
- Camera systems and image sensors.
- Synthetic data generation and domain randomization.
- 3D graphics, Blender, Unreal Engine, or similar tools.
- Self-supervised, semi-supervised, or active learning.
- Signal processing, mathematics, optimization, or state estimation.
- Virtual reality, robotics, tracking systems, or motion capture.
- Personal projects, research work, open-source contributions, or a strong technical portfolio.
- A maker mindset: building prototypes, testing ideas, learning from failures, and iterating quickly.
Academic credentials are not required. Practical implementation ability is more important than academic credentials.
What We Offer
- Work on machine learning technology that will operate directly inside real consumer hardware products.
- Involvement in new R&D projects from early prototypes through to production.
- Technically challenging work involving machine learning, embedded systems, electronics, tracking, virtual reality, and robotics.
- Significant autonomy and direct influence over technical decisions.
- Access to our own laboratory, GPUs, development hardware, prototypes, test equipment, and production tools.
- Flexible working hours.
- Possibility of partial remote work when compatible with the role.
- Competitive salary based on experience and technical ability.
- Salary reviews linked to responsibility, performance, and impact.
