Roel H.

• Rådgivningsarbeid innen dataanalyse og krypto.

Clients and Projects:

- Electrolux is a Swedish multinational home appliance manufacturer headquartered in Stockholm.

• Utilized Python packages (lifetimes, attribution, surpyval) to analyze customer purchase behaviors and marketing channel effectiveness.
• Optimized marketing strategies and the sales funnel through detailed customer journey mapping, increasing conversions and customer satisfaction.

...

- BEF4U is a Belgium company predicting the energy consumption in shopping malls

• Developed predictive models to forecast energy consumption in shopping malls.
• Analyzed large datasets to identify patterns and optimize energy usage, contributing to cost-saving strategies.

...

- Uptime water management provider

• Developed a Python-based system comprising a dashboard for visualization and parameter tuning, and a batch processing engine for analyzing sensor data.
• The AI model learned to identify pump usage by detecting temperature changes when water was pumped.

...

- Polpo, LLM project, using a Dutch model and Langchain to process big documents in chunks (political papers, reports of discussions from Dutch parliament). Found the right model based on language and confidentiality needs. Worked on minimizing the response time.

...

- fBeta GmbH, evaluation of LLM generated answers versus human expert answers

• different evaluation metrices were developped to compare human expert answers verus LLM generated answers based on word embeddings.

• Bidratt til å utvikle verdens første sensorbaserte poengsystem for den profesjonelle Bull Riders-organisasjonen ved hjelp av kunstig intelligens.
• Ansvarlig for implementering av analyseløsningen for poengsystemet.
• Ferdigstilt prosjekt for kundetilfredshet.
• Deltatt i ulike førsalgsaktiviteter, konseptbevis omfangsdefinisjon for flere prosjekter innen sport og underholdning, inkludert basketball, ridning og surfing.

AGT International is a leader in Artificial Intelligence for IoT solutions. Proven expertise is across various sectors, including smart cities, manufacturing, safety and security, and enabling independent living for the elderly.

• Provided a major contribution to developing the world's first sensor-based bull scoring system using artificial intelligence for the Professional Bull Riders organization.

• Responsible for the implementation of the analytics solution of bull scoring. This included sensor signal processing, feature extraction of bull parameters, and model building to predict the difficulty of a ride according to multiple criteria.

• Finalized project to customer satisfaction.

• Participated in various pre-sales activities, proofs-of-concept and scope definition for multiple projects in sports and entertainment, including basketball, horse riding, and surfing.

• Bygget prototyper og produksjonskode for ulike bruksområder i industri 4,0, inkludert tilstandsbasert vedlikehold for tungt maskineri, beregning av maskintilstand og deteksjon av avvikelser.
• Bidratt til et system for grøftstabilitetsovervåking med in-situ-sensorer Analysearbeidet omfattet feilforutsigelse, trendanalyse, modellering og simulering av komplekse miljøsystemer.

AGT International is a leader in Artificial Intelligence for IoT solutions. Proven expertise is across various sectors, including smart cities, manufacturing, safety and security, and enabling independent living for the elderly.

• Build prototypes and production code for various applications in Industry 4.0, including predictive maintenance for heavy machinery, machine state estimation, and anomaly detection. Customers were from the German industry.

• Contributed to a system for dike stability monitoring with in-site sensors.

• The analytics work included failure prediction, trend analysis, modeling, and simulation of complex environmental systems.

• Developed analytics code for various projects, including automatic flood detection using radar imagery, non-destructive testing of material defects using Terahertz radiation, mine detection on the seafloor using synthetic aperture sonar imagery