Jens Henriksson Email & Phone Number

Gothenburg, Sweden

Working as a Senior Data Scientist focusing on designing and implementing data and AI workflows for our clients. Work includes incorporating computer vision models ( e.g., Semantic segmentation, Object Detection) into automation solutions and language processing models (e.g., transformer based large language models) into workflows to transform or extract information from data sources. In addition, leading an AI and Data community that celebrates 10 years in 2025. The focus of the community has been to stay up to speed with AI news, including reading and analyzing research publications, as well as sharing tips-and-tricks on how to utilze the latest trends in AI.

Gothenburg, Sweden

Industrial Phd focusing on how outlier detection can be utilized as a safety measure towards safety critical application. The thesis bridged the gap between Safety Engineers and Deep Learning Engineers by formulating metrics that explains risk and performance in such a way that they are suitable to both parties.Successfully defended the thesis October 2023, with a total of 10 peer reviewed publications.

Investigated how automation can be incorporated and recommended for road construction sites. The project focused on how digitalization, electrification and automation will benefit road constructions in their journey towards no local emissions, sustainable energy sources and less than 4 accidents per 1 million work hours.

Industrial Phd focusing on how outlier detection can be utilized as a safety measure towards safety critical application. The thesis bridged the gap between Safety Engineers and Deep Learning Engineers by formulating metrics that explains risk and performance in such a way that they are suitable to both parties.Successfully defended the thesis October 2023, with a total of 10 peer reviewed publications.

Developed supervision strategies for machine learning models in the SMILE II and SMILE III projects, implementing two safety cage approaches that include end-to-end variational autoencoders and anomaly detection for an autonomous driving simulation.

Led the development of a Python-based federated learning toolbox, optimizing for in-vehicle hardware integration while managing team execution and technical implementation.

Managed a team of ML Engineers that constructed and merged various sources of datasets that was the base for training semantic segmentation DNNs of road environments.

Implemented image analysis pipelines and CNN-based object detection systems for robotic lawnmower navigation

Developed and implemented state-of-the-art 3D CNN classification models operating on LIDAR data, as well as managing cross-team collaboration between Semcon SE and Semcon Brazil.

Gothenburg, Sweden

The Master thesis evaluated if it were possible to determine when splash and spray from adjacent vehicles were happening. The goal was to create smart algorithms that would improve the current wiper system to react in a proactive manner, in order to keep a clear view, without creating any false strokes. During the thesis, a splash- and spray model were developed. The model used information from the vehicles' camera and radar, especially adjacent vehicles positions and velocities, to evaluate if any of the surrounding vehicle would cause splash or spray. In addition, the information were used in a validation process to estimate the water film thickness on the road surface. The estimation were done using a Kalman Filter, that got measurement updates with the nonlinear splash model. The Thesis gave good results, especially in the highway scenario, but also decent results on country roads when heavy vehicles occurred from time to time. Algorithms were developed in Matlab/Simulink, and used both pre-gathered data and own gathered. The thesis was recently showcased in a marketing campaign at Semcon. You can see me talking about the solution in the video here: https://www.youtube.com/watch?v=gzqdNvfBGzg

Torslanda

Assisted during vacation periods 2011, 2012 and 2014. During the first years I remade most of the work instructions for the four production lines. This included doing some time analysis, rewriting work instructions in a uniform format. In addition, some changes to line-balancing were made, and also FMEA. In addition, the production layout were also heavily updated (made in AutoCAD), to fit reality. During the last year I got the task to create a pre-kitting area for the upcoming production line. This included a full out engineering solution of material flow, layout and worker balance. The area were sketched, planned, created, and finally tested in live production. This included supervision of operators as well as teaching how to use the new system. After the system proved successful, additional improvements were made, using AGV's to transfer the line-kitts to the actual line.

Stenungsund

Worked during the summer. Were part of a four-man team that served costumers during the evenings, by preparing products, helping as a cashier and delivering postal packages.

Phd, Computer Science

The field of deep learning has elevated computer vision to new heights compared to traditional image processing methods. Incorporating.

Master Of Science (M.Sc.), Systems, Control And Mechatronics (Mpsys)

Bachelor’S Degree, Automation Och Mekatronik (Z)

Activities and Societies: (2011) Z-sektionens Mottagningskommitté (2011-2012) Z-sektionens Valberedning (2012-2013) Ordförande för.