Matthijs Brouns

Xccelerated is a part of the Xebia Group, accelerating careers and skills of professionals. Through our 13-month program, we train highly skilled Machine learning engineers, Data engineers & Cloud professionals with the goal to become authorities in their fields and be connected to their next employer.

As Xccelerated’s CTO, I:

• Led the change of our business model to be more in line with the market trends during and after COVID. This allowed us to grow the organisation from 15 FTE to 70 FTE within a year;
• Set up 5 new technical competencies (MLOps, Cloud Engineering, DevOps Engineering, Analytics Engineering and FinOps) employing 55 FTE
• Build and train teams at ASML to shape their cloud cost management practice;
• Consulted on the MLOps competency at ASML by designing core aspects of the system as well as by training key stakeholders and users of the system;
• Built software at ASML to move their software testing workload to Google Cloud Platform;
• Kickstarted Heineken’s Global Data & Analytics team by seeding the team with capable consultants, training key internal Heineken domain experts on Data & AI and assisting them through developing various MVP solutions;
• Designed, Built and delivered a VAT engine at bol.com in Google Cloud Platform;
• Facilitated a data-driven strategy at Novartis’ translational medicine department through stakeholder sessions, training and workshops;
• Designed, consulted on, and led teams to implement Data Platforms & Warehouses at clients like Allianz, Woonstad Rotterdam, Colliers and others;

As Machine Learning Engineer at Riviq I act as an advisor and lead developer in big data and analytics projects. At Riviq I worked on among others:

• Deduplicated scraped vacancy texts on large scale using a Siamese Deep Learning architecture and custom-built approximate nearest neighbour algorithms;
• Performed Entity Extraction on vacancy texts using Attention based Deep learning models in order to facilitate faceted search
• Built a framework to easily run Machine Learning models in AWS Sagemaker
• An online vacancy recommender using Relay42 clickstream data with AWS Kinesis and Annoy
• Built a serverless data lake on AWS using Athena, S3 and Lambda
• Deployed and Maintained an Airflow cluster on an ECS cluster
• Estimated labour market tension for different occupational areas using PyMC3
• Predicted churn for SME customers to improve retention
• Designed and built a Personal Information vault solution which is able to maintain data as an immutable log while being GDPR-compliant.

The software development department at Systems Navigator was traditionally a relatively small and unstructured team of two full-time developers. However, with several major opportunities for growth, there was a need to increase the productivity of the department. As Software Development Manager at Systems Navigator, I:

• Designed and built two event-based real-time data integration solutions between clients’ ERP systems and our in-house built planning solutions
• Built and deployed several Machine Learning models that facilitated planning and scheduling of liquid bulk terminals
• Implemented a mature DevOps workflow
• Managed and extended a team of 6 software developers.
• Increased the percentage of projects on time from 40% to 85%.
• Increased average client satisfaction by 2 points out of 10.
• Increased codebase quality by introducing code reviews and individual educational tracks.

Because of these changes, the software development department was able to have its first profitable year. Apart from managing the team I remained active as a developer implementing various microservices for predictive analytics and scheduling algorithms at liquid bulk terminals

Systems Navigator is an independent software consultancy firm based in Delft, The Netherlands. It specializes in the design and creation of decision support solutions based on Operations Research and machine learning technology. Within Systems Navigator I, among others:

• Designed and build a scheduling support system for application in liquid bulk storage terminals.
• Developed predictive maintenance models for liquid bulk terminals.
• Designed and implemented several web-based operational planning support solutions based on direct feedback and easy user interaction at a.o. the Panama Canal and LNG terminals.
• Developed a scheduling system for infrastructure planning at liquid bulk terminals.
• Maintained and developed a Simulation as a Service Web platform, extending it with functionality for a.o. experimental design and analysing large numbers of experiments.
• Designed, built, and tested discrete-event simulation models of liquid bulk terminals and refineries.