Applied AI for mid-sized companies
The full path of an AI project: clarify feasibility, build the model, take it into production.
Co-Founder of myBytes
Guido Winger
Software engineer and co-founder of myBytes.
Earlier in his career, he developed and led the e-commerce platform xt:Commerce. Today he translates the business problems of mid-sized companies into AI models that run in production, from the data audit to the business case in EUR.
Applied AI & Data Science
Current focus
Four areas of work: two lead to published case studies, one to the research section, one to the consulting offer.
Forecasting and commodity intelligence
Forecast and regime models for commodity and demand data, documented in the research section.
Recommender systems
Recommendation systems with transparent methodology, including published code.
Deep learning
Neural networks where they measurably outperform classical methods, for example in image recognition.
Guido Winger and xt:Commerce
Engineering history
Guido Winger co-developed xt:Commerce, carried responsibility for the system early on and served as managing director of xt:Commerce GmbH until 2010. Academic papers published in 2005 and 2006 explicitly credit him and Mario Zanier with developing the shop software.
International trademark registers, including national records in Canada, Australia and the United States, identify Guido Winger personally as the holder of the historical xt:Commerce word/device mark and also document the later assignment of the mark to xt:Commerce GmbH.
During his years with the project, xt:Commerce achieved substantial international adoption. In 2008, Heise reported that xt:Commerce was running on an estimated 100,000 servers worldwide, and the 2008 publication by ibi research at the University of Regensburg presented xt:Commerce as one of the leading open-source e-commerce solutions. The June 2010 acquisition announcement listed reference customers including Osram, Volkswagen Accessories, 3M, Butlers, Mindfactory, Marktkauf, Delsey and Bugatti.
Independent evidence
This history is not self-reported. It can be verified against primary sources:
- WIPO Gazette 2007: International Registration No. 920332, holder Guido Winger
- Canadian trademark register: applicant and owner Guido Winger, assignment in 2010
- USPTO register: US trademark XT:COMMERCE (Reg. No. 3402105), registrant Guido Winger
- IP Australia official journal, 2007 (PDF): Australian registration No. 1176841, owner Guido Winger
- xt:Commerce GmbH emerged from the Winger & Zanier partnership (GbR), documented in court (Düsseldorf Higher Regional Court, 2010, I-20 U 41/09)
- Heise online, 2008 (report on a payment-system security flaw): cites an estimated 100,000 xt:Commerce servers worldwide
- c't 6/2008, page 96 (archived PDF): print version of the Heise report
- Diploma thesis, VWA Wiesbaden 2006 (PDF): names Mario Zanier and Guido Winger as the creators of xt:Commerce
- ibi research at the University of Regensburg, 2008: xt:Commerce presented as one of the leading open-source e-commerce solutions
- Acquisition announcement, June 2010 (via Finextra): over 100,000 shops and the reference customer list
Research by Guido Winger
Research
Selected work with disclosed methodology and code.
Malaria detection with deep learning
MIT capstone project: CNN classification of blood cell images, fully reproducible.
Amazon product recommendations
Recommender systems compared, from popularity baselines to matrix factorisation.
Building an AI web application
What AI-assisted development can handle today and where it breaks down, documented in a real project.
Open source and code
Code
Public repositories containing Guido Winger's work, in the myBytes research organisation on GitHub.
Certifications
Credentials
Exact programme names with the supporting proof.
Applied AI and Data Science Program
MIT Professional Education
Deep Learning Specialist
Bitkom Akademie
Professional Scrum Master
Scrum.org, verified via Credly
Work with Guido Winger
Which business problem should a model solve first? An initial conversation clarifies feasibility and data readiness, with no sales deck.