Machine Learning Engineer

IntroductionDeCentralCode:

DeCentralCodeis a technology company that applies modern solutions to address current challenges. With locations in Rotterdam, Netherlands and Salem, India, weoperateon a global scale.

Ourexpertiselies incutting-edgetechnologies such as machine learning/artificial intelligence, distributed ledger technology, blockchain, Internet of Things, cloud-native, serverless architecture, and cybersecurity.

We collaborate with knowledge institutions, governments, and businesses to co-create andvalidatesolutions for various domains including supply chain and energy transition, among others.

As an early-stage partner, we offer the opportunity to work with thecutting-edgetechnologiesfromscratch.

Location:

Salem, Tamil Nadu, India.

We are looking for:

Were looking for an ML/AI Engineer to own the complete machine learning pipeline for a medical imaging application from dataset curation through model deployment. Youll build classification systems, implement domain adaptation for real-world robustness, and create statistical validation infrastructure.

Ability toidentifyand solve complex problems using machine learning techniques, with a focus on delivering actionable insights and tangible business value.

Implement best practices for testing and validation of machine learning models, ensuring robustness, reliability, and reproducibility of results.

Effectively communicate technical concepts and insights to non-technical stakeholders, and present findings and recommendations in a clear and concise manner.

Responsibilities:

• Own the end-to-end machine learning pipeline, from data ingestion to model evaluation.
• Build and maintain dataset pipelines with support for multi-annotator workflows.
• Implement data augmentation strategies to improve model generalisation.
• Ensure proper train / validation / test splits for reliable model performance.
• Develop CNN-based multi-output ordinal classification models using architectures such as ResNet and EfficientNet.
• Optimise models for mobile and edge deployment using ONNX and TensorRT.
• Implement domain adaptation techniques to ensure robustness across:
• Different devices
• Varying lighting conditions
• Diverse user environments
• Build image quality assessment modules to filter or score input data.
• Implement uncertainty quantification to measure prediction confidence.
• Apply probability calibration techniques for reliable output probabilities.
• Set up statistical evaluation frameworks with:
• Reproducible experiments
• Consistent validation pipelines
• Reliable performance metrics

Necessary Skills:

Must Have:

• 4+ years of hands-on experience in deep learning and computer vision.
• StrongproficiencyinPyTorchand deep learning workflows.
• Solid experience with CNN architectures such asResNet,EfficientNet, or equivalent.
• Good understanding of ordinal classification and multi-output learning techniques.
• Expertisein data pipeline development, including preprocessing, augmentation, and validation.
• Strong knowledge of statistical analysis and model evaluation methodologies.
• Proficiencyin Python and its data science ecosystem: NumPy, pandas, scikit-learn, matplotlib/seaborn.
• Experience with Git version control and clean, maintainable coding practices.

Good to Have:

• Experience working with medical imaging or healthcare AI solutions.
• Knowledge of mobile and edge model optimization techniques (ONNX,TensorRT,CoreML).
• Familiarity with domain adaptation and transfer learning methods.
• Exposure to Bayesian deep learning and uncertainty quantification approaches.
• Experience using data annotation tools and managing annotation workflows.