Erwin Rudi

ML Engineer with 4 years of experience building and shipping production ML systems across the full lifecycle - from model development and training to pipeline design and cloud deployment. I've worked across computer vision, NLP, data engineering and cloud infrastructure, most recently building an AI-powered computer vision product that has analysed over 100 million images in production, alongside the underlying ML platform on AWS EKS. I enjoy working at the intersection of ML engineering and cloud platform architecture, and thrive as a hands-on engineer in collaborative teams with strong technical depth.

Machine Learning Engineer at EnaoVision (2025-06 – Present)

• Built and maintained the ML Platform infrastructure on AWS (EKS, RDS, VPCs, Load Balancers, Monitoring) fully managed through Terraform.
• Co-developed the core ML system powering AI-based quality control for manufacturers including BMW and Birkenstock, contributing to over 100 million products inspected to date.
• Designed and optimised three production ML pipelines covering segmentation, anomaly detection and error classification - with a focus on parallelization, resource efficiency and throughput.
• Introduced multi-GPU support on ML platform, achieving 8× faster model training and inference.
• Implemented checkpoint-based fault tolerance for training jobs, enabling use of spot instances and significantly reducing cloud compute costs.

Cloud Engineer at EnaoVision (2024-12 – 2025-05)

• Part of a 3-person MLOps team owning infrastructure and automation across 10 concurrent client projects on AWS and Azure.
• Managed and refined CI/CD workflows using Terraform, Azure DevOps and AWS CodePipeline; integrated cloud services including storage, databases, Kubernetes clusters and containerised applications.
• Developed two ETL pipelines with Flyte on AKS: one ingesting and aggregating marketing data from Instagram and LinkedIn APIs into a PowerBI dashboard via Postgres; one processing industrial extrusion machine data for operational reporting.
• Built a RAG knowledge assistant prototype for internal documents on AWS Fargate, using Bedrock for inference; implemented chunking, embedding, and RAGAS-based evaluation of retrieval quality.

Machine Learning Engineer / Data Engineer at Alexander Thamm GmbH (2022-06 – 2024-05)

• Developed an intent detection pipeline for customer support emails using FastAPI and ML pipelines, automating ∼ 10k emails per week.
• Built two ETL pipelines on an Azure analytics platform (medallion architecture, ADLS, Synapse): one migrating on-premise HR data via PySpark for PowerBI reporting; one extracting competitor products and pricing from scanned journals using OCR.
• Reimplemented a research paper in PyTorch as part of an autonomous driving initiative. Trained and evaluated computer vision models.
• Built an internal full-stack app to visualise personalised learning paths for data engineers across the team.
• Served as internal and external trainer for Python and software engineering best practices.

Data Scientist - Working Student at RWTH Aachen University - Physics Institute 3a (2019-12 – 2021-11)

• Supported migration from VMs to Docker containers, improving environment reproducibility for the research team.
• Built a user management system in Vue.js as part of a full-stack web application for the institute.
• Trained deep learning models for a particle physics experiment at CERN using TensorFlow on a GPU cluster.

M.Sc. in Physics – RWTH Aachen University (2019-10 – 2021-11)

B.Sc. in Physics – RWTH Aachen University (2016-10 – 2019-09)