Looks like you have JavaScript disabled. For the full Twine experience, you will need to re-enable it.

I’m an ML engineer who specializes in building models that operate under real-world constraints: limited data, noisy signals, and imperfect measurements. My background is in data science and applied machine learning, with hands-on experience in generative models, image and signal processing, and anomaly detection. I’ve worked on both research-driven and industrial machine learning projects, ranging from diffusion-based generative models for image reconstruction to production-oriented anomaly detection systems using acoustic data. I’m comfortable taking problems end-to-end: framing them correctly, designing or selecting models, running large-scale experiments, and translating results into usable systems. What differentiates me from other developers is a strong focus on modeling under uncertainty. I’ve worked on inverse problems where multiple solutions are possible and data is sparse or noisy by default. This has made me rigorous about assumptions, evaluation, and failure modes—skills that carry directly into industrial ML systems. My technical stack centers on Python (PyTorch, scikit-learn, Pandas), with experience in high-performance and multi-GPU environments. My sector experience includes generative AI, scientific image reconstruction, and industrial anomaly detection. --- Employment and project experience Research Intern – Image Reconstruction & Generative AI _September 2025 – December 2025_ Worked on image reconstruction from sparse and noisy measurements (inverse problems) where multiple plausible solutions exist. Designed and implemented diffusion-based generative models to sample measurement-consistent reconstructions, achieving large improvements over classical baselines. Conducted extensive experimental evaluations; results were submitted for peer review. Machine Learning Intern – Industrial Anomaly Detection (Rolex) _September 2024 – February 2025_ Developed machine learning models for acoustic anomaly detection and machine state classification using industrial sound data. Built unsupervised detection systems and supervised classifiers under limited labeling constraints. Evaluated robustness and feasibility using both internal data and targeted open-source datasets.…I’m an ML engineer who specializes in building models that operate under real-world constraints: limited data, noisy signals, and imperfect measurements. My background is in data science and applied machine learning, with hands-on experience in generative models, image and signal processing, and anomaly detection. I’ve worked on both research-driven and industrial machine learning projects, ranging from diffusion-based generative models for image reconstruction to production-oriented anomaly detection systems using acoustic data. I’m comfortable taking problems end-to-end: framing them correctly, designing or selecting models, running large-scale experiments, and translating results into usable systems. What differentiates me from other developers is a strong focus on modeling under uncertainty. I’ve worked on inverse problems where multiple solutions are possible and data is sparse or noisy by default. This has made me rigorous about assumptions, evaluation, and failure modes—skills that carry directly into industrial ML systems. My technical stack centers on Python (PyTorch, scikit-learn, Pandas), with experience in high-performance and multi-GPU environments. My sector experience includes generative AI, scientific image reconstruction, and industrial anomaly detection. --- Employment and project experience Research Intern – Image Reconstruction & Generative AI _September 2025 – December 2025_ Worked on image reconstruction from sparse and noisy measurements (inverse problems) where multiple plausible solutions exist. Designed and implemented diffusion-based generative models to sample measurement-consistent reconstructions, achieving large improvements over classical baselines. Conducted extensive experimental evaluations; results were submitted for peer review. Machine Learning Intern – Industrial Anomaly Detection (Rolex) _September 2024 – February 2025_ Developed machine learning models for acoustic anomaly detection and machine state classification using industrial sound data. Built unsupervised detection systems and supervised classifiers under limited labeling constraints. Evaluated robustness and feasibility using both internal data and targeted open-source datasets.

Michel Morales

AI Engineer, AI Developer, Data Scientist, +2





I’m an ML engineer who specializes in building models that operate under real-world constraints: limited data, noisy signals, and imperfect measurements. My background is in data science and applied machine learning, with hands-on experience in generative models, image and signal processing, and anomaly detection. I’ve worked on both research-driven and industrial machine learning projects, ranging from diffusion-based generative models for image reconstruction to production-oriented anomaly detection systems using acoustic data. I’m comfortable taking problems end-to-end: framing them correctly, designing or selecting models, running large-scale experiments, and translating results into usable systems. What differentiates me from other developers is a strong focus on modeling under uncertainty. I’ve worked on inverse problems where multiple solutions are possible and data is sparse or noisy by default. This has made me rigorous about assumptions, evaluation, and failure modes—skills that carry directly into industrial ML systems. My technical stack centers on Python (PyTorch, scikit-learn, Pandas), with experience in high-performance and multi-GPU environments. My sector experience includes generative AI, scientific image reconstruction, and industrial anomaly detection. --- Employment and project experience Research Intern – Image Reconstruction & Generative AI _September 2025 – December 2025_ Worked on image reconstruction from sparse and noisy measurements (inverse problems) where multiple plausible solutions exist. Designed and implemented diffusion-based generative models to sample measurement-consistent reconstructions, achieving large improvements over classical baselines. Conducted extensive experimental evaluations; results were submitted for peer review. Machine Learning Intern – Industrial Anomaly Detection (Rolex) _September 2024 – February 2025_ Developed machine learning models for acoustic anomaly detection and machine state classification using industrial sound data. Built unsupervised detection systems and supervised classifiers under limited labeling constraints. Evaluated robustness and feasibility using both internal data and targeted open-source datasets.…I’m an ML engineer who specializes in building models that operate under real-world constraints: limited data, noisy signals, and imperfect measurements. My background is in data science and applied machine learning, with hands-on experience in generative models, image and signal processing, and anomaly detection. I’ve worked on both research-driven and industrial machine learning projects, ranging from diffusion-based generative models for image reconstruction to production-oriented anomaly detection systems using acoustic data. I’m comfortable taking problems end-to-end: framing them correctly, designing or selecting models, running large-scale experiments, and translating results into usable systems. What differentiates me from other developers is a strong focus on modeling under uncertainty. I’ve worked on inverse problems where multiple solutions are possible and data is sparse or noisy by default. This has made me rigorous about assumptions, evaluation, and failure modes—skills that carry directly into industrial ML systems. My technical stack centers on Python (PyTorch, scikit-learn, Pandas), with experience in high-performance and multi-GPU environments. My sector experience includes generative AI, scientific image reconstruction, and industrial anomaly detection. --- Employment and project experience Research Intern – Image Reconstruction & Generative AI _September 2025 – December 2025_ Worked on image reconstruction from sparse and noisy measurements (inverse problems) where multiple plausible solutions exist. Designed and implemented diffusion-based generative models to sample measurement-consistent reconstructions, achieving large improvements over classical baselines. Conducted extensive experimental evaluations; results were submitted for peer review. Machine Learning Intern – Industrial Anomaly Detection (Rolex) _September 2024 – February 2025_ Developed machine learning models for acoustic anomaly detection and machine state classification using industrial sound data. Built unsupervised detection systems and supervised classifiers under limited labeling constraints. Evaluated robustness and feasibility using both internal data and targeted open-source datasets.

Available to hire

I’m an ML engineer who specializes in building models that operate under real-world constraints: limited data, noisy signals, and imperfect measurements. My background is in data science and applied machine learning, with hands-on experience in generative models, image and signal processing, and anomaly detection.

I’ve worked on both research-driven and industrial machine learning projects, ranging from diffusion-based generative models for image reconstruction to production-oriented anomaly detection systems using acoustic data. I’m comfortable taking problems end-to-end: framing them correctly, designing or selecting models, running large-scale experiments, and translating results into usable systems.

What differentiates me from other developers is a strong focus on modeling under uncertainty. I’ve worked on inverse problems where multiple solutions are possible and data is sparse or noisy by default. This has made me rigorous about assumptions, evaluation, and failure modes—skills that carry directly into industrial ML systems.

My technical stack centers on Python (PyTorch, scikit-learn, Pandas), with experience in high-performance and multi-GPU environments. My sector experience includes generative AI, scientific image reconstruction, and industrial anomaly detection.

Employment and project experience

Research Intern – Image Reconstruction & Generative AI
September 2025 – December 2025

Worked on image reconstruction from sparse and noisy measurements (inverse problems) where multiple plausible solutions exist.
Designed and implemented diffusion-based generative models to sample measurement-consistent reconstructions, achieving large improvements over classical baselines.
Conducted extensive experimental evaluations; results were submitted for peer review.

Machine Learning Intern – Industrial Anomaly Detection (Rolex)
September 2024 – February 2025

Developed machine learning models for acoustic anomaly detection and machine state classification using industrial sound data.
Built unsupervised detection systems and supervised classifiers under limited labeling constraints.
Evaluated robustness and feasibility using both internal data and targeted open-source datasets.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Language

French

Fluent

English

Fluent

German

Advanced

Work Experience

Research Intern - Image Reconstruction / Generative AI

September 1, 2025 - December 1, 2025

Worked on image reconstruction from sparse and noisy measurements (inverse problems) where multiple plausible restorations exist. Developed a diffusion-based generative model to sample measurement-consistent image reconstructions. Achieved up to 3 orders of magnitude improvement vs. baseline methods on key evaluation metrics. Strengthened the approach through extensive experiments; results submitted for peer review.

Machine Learning Intern at Rolex

September 1, 2024 - February 1, 2025

Developed machine learning models for acoustic anomaly detection and machine state classification using industrial sound data. Built unsupervised models to detect failing sources and supervised models to classify operational states. Addressed limited labeled data by augmenting experiments with targeted open-source datasets and conducting feasibility evaluations under data constraints.