Herman Groenbroek
As a full-stack AI Engineer with a passion for creativity, I love to design and build innovative
PoCs using the state-of-the-art in AI. I develop practical end-to-end solutions in Python, and I am
well-versed in DevOps and Agile Scrum methods. My strengths include effective communication and a
strong desire to learn and share knowledge. When it comes to developing intelligent systems, you can
count on me to deliver top-notch results.
Certified PSM-I | AZ-900 | AI-900.
Nationaal Coördinator Groningen via Sopra Steria
10-2024 – present
Senior Business Consultant at the AI Center of Excellence at Nationaal Coördinator Groningen (NCG). Led the initial strategic assessment of AI use case opportunities across the organization, identifying and recommending prioritisation of projects aligned with business goals with impact/effort matrices. Key contributor to our AI solution architecture for future, human-centered AI implementations, to ensure scalability and modularity. Designed, developed and deployed an AI assistant on Azure focusing on improving employee productivity, providing users with LLM-based applications for document summarization and language translation to B1-level, as well as being a privacy-centered alternative to ChatGPT aligned with the GDPR and EU AI Act. Taught departments about the potential of AI at NCG via workshops and presentations.
Sopra Steria
07-2024 – present
Data Scientist internally at Sopra Steria (formerly Ordina). Played a key role in launching a new Data & AI practice in the north, including the creation of a comprehensive Development Plan detailing the skills progression for Data Scientists at all levels. I designed and delivered Prompt Engineering workshops to colleagues, fostering AI literacy and promoting responsible AI.
CSIRT-DSP via Sogeti
05-2023 – 05-2024
Infrastructure Engineer at the national Computer Security Incident Response Team (CSIRT-DSP), part of the Ministry of Economic Affairs. I played a key role in designing, developing, and maintaining infrastructure components, including data collection and automation tools built as IntelMQ bots, containerized Python applications using Docker and supporting APIs. My work contributed to the capabilities of the Digital Trust Center (DTC) as well, enhancing the Netherlands’ cybersecurity posture and resilience.
Provincie Drenthe via Sogeti
04-2023 – 07-2023
Deployment Coordinator overseeing the rollout and implementation of devices to Members of the States. Responsibilities included strategic rollout planning, implementation optimization and thorough testing to ensure a smooth experience for all. Delivered extensive training sessions to Members of the States on device features and Microsoft 365 applications. Provided ongoing technical support and issue resolution via TOPdesk for remaining aid, which was rarely needed in practice.
Sogeti
06-2022 – 06-2024
Data Scientist at Sogeti. Taught about AI and upcoming Large Language Models. Scrum Master and internship mentor for the Brain-Computer Interface WBSO project, where we developed an ensemble neural network using feature engineering methods on EEG motor cortex signals in order to predict actions in a game, with the end goal of empowering disabled people to regain control.
Klippa
11-2021 – 05-2022
AI Engineer at Klippa. Developed and deployed improvements to intelligent document processing algorithms, enhancing ID document recognition accuracy and incorporating address parsing capabilities. Demonstrated commitment to code quality and maintainability through code reviews, documentation, and adherence to coding standards. Integrated data from external APIs and leveraged NLP with Flair for advanced text analysis. Trained and deployed YOLO Computer Vision models for object detection within the core document intelligence pipeline.
Slimmer AI
05-2019 – 04-2021
Graduation Internship at Slimmer AI. Developed a novel dataset—the 6L5K Music Corpus—and applied machine learning techniques for Automatic Language Identification of vocals in music solely from audio waveform data. This was achieved by developing a vocal detector, turning vocal fragments into mel spectrograms and MFCCs, and classifying the transformed vocal fragments as one of the six languages used by various VGGish trainings, a deep neural network, and an ensemble of the two.
2017 – 2021
MSc. Artificial Intelligence
Specialized in Computational Intelligence & Robotics. Average grade: 7.6.
Master's Thesis: "A Machine Learning approach to Automatic Language Identification of vocals in music".
University of Groningen
2013 – 2017
BSc. Artificial Intelligence
Bachelor's Thesis: "Improving natural image denoising using a multilayer perceptron".
University of Groningen