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

I am Michael Woelki, a Machine Learning Software Engineer specializing in architecting scalable data pipelines and training AI agents. I have 4+ years of experience building high-performance Python systems and deep learning research, with hands-on work in PyTorch, JAX, and transformer-based models. I enjoy turning complex data into robust models, auditing model reasoning, and designing reproducible experiments. Outside work, I contribute to Kaggle competitions and help drive AI safety and tooling improvements through engineering best practices.…I am Michael Woelki, a Machine Learning Software Engineer specializing in architecting scalable data pipelines and training AI agents. I have 4+ years of experience building high-performance Python systems and deep learning research, with hands-on work in PyTorch, JAX, and transformer-based models. I enjoy turning complex data into robust models, auditing model reasoning, and designing reproducible experiments. Outside work, I contribute to Kaggle competitions and help drive AI safety and tooling improvements through engineering best practices.

Michael Woelki

AI Engineer, Data Scientist, Full Stack Developer, +2





I am Michael Woelki, a Machine Learning Software Engineer specializing in architecting scalable data pipelines and training AI agents. I have 4+ years of experience building high-performance Python systems and deep learning research, with hands-on work in PyTorch, JAX, and transformer-based models. I enjoy turning complex data into robust models, auditing model reasoning, and designing reproducible experiments. Outside work, I contribute to Kaggle competitions and help drive AI safety and tooling improvements through engineering best practices.…I am Michael Woelki, a Machine Learning Software Engineer specializing in architecting scalable data pipelines and training AI agents. I have 4+ years of experience building high-performance Python systems and deep learning research, with hands-on work in PyTorch, JAX, and transformer-based models. I enjoy turning complex data into robust models, auditing model reasoning, and designing reproducible experiments. Outside work, I contribute to Kaggle competitions and help drive AI safety and tooling improvements through engineering best practices.

Available to hire

I am Michael Woelki, a Machine Learning Software Engineer specializing in architecting scalable data pipelines and training AI agents. I have 4+ years of experience building high-performance Python systems and deep learning research, with hands-on work in PyTorch, JAX, and transformer-based models.

I enjoy turning complex data into robust models, auditing model reasoning, and designing reproducible experiments. Outside work, I contribute to Kaggle competitions and help drive AI safety and tooling improvements through engineering best practices.

Skills

Experience Level

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Expert

Language

Bashkir

Intermediate

Javanese

Advanced

Work Experience

Machine Learning Engineer (Contract) at Applied AI Training

January 1, 2024 - Present

Designing multi-step ML challenges (gradient stability, custom CUDA kernels, distributed training) for AI agents to solve in isolated Python environments. Audited failure modes in code-generation models, identifying bottlenecks in JAX-based XLA compilation and PyTorch autograd. Developed standardized tasks for high-dimensional optimization and neural architecture search. Curated high-quality golden responses for supervised fine-tuning of frontier coding models.

ML Research Engineer at Fraunhofer Institute (IIS)

September 1, 2021 - December 1, 2023

Developed end-to-end Python pipelines for processing large-scale sensory data; designed Transformer-based architectures for time-series forecasting using PyTorch; tracked experiments with Weights & Biases; maintained robust Git-based CI/CD workflows with 95%+ test coverage for production-level ML modules.