Patrick Leuchtenberger

Topic: Drift-aware Architecture for Adaptive Reinforcement Learning in Semiconductor Production Control

- Improved manufacturing quality and repeatability, measured by ISO test results through optimization of CNC machining and bonding processes - Reduced prototype iteration time, measured by turnaround speed by designing assemblies in PTC Creo and producing functional prototypes via FDM and SLA printing - Validated production parameters, measured by sensor and inspection data using pressure loggers, optical microscopy and standardized functional testing - Stabilized chip module integration, measured by bonding yield and reliability by planning and implementing chip bonding processes (ACF, flip-chip, T-connect) including pre- and post-processing of chip modules - Enabled scalable smartcard production, measured by process robustness through end-to-end manufacturing planning from card body construction to chip module preparation and bonding

- Established a validated cardiovascular testbed, measured by physiological realism by translating anatomical components into a controlled scientific testing environment - Enabled MRI-compatible flow experiments, measured by data consistency through material, sensor, and fluid selection for 4D flow MRI - Validated hemodynamic models, measured by agreement between simulation, sensors and MRI data using pressure and flow measurements

- Enabled robotic navigation accuracy, measured by system stability by processing LiDAR sensor data in ROS - Integrated SLAM algorithms, measured by localization robustness into a navigation system for a medical C-arm - Bridged engineering and clinical practice by participating in laparoscopic procedures to validate real-world constraints

- Accelerated material and component validation measured by reduced evaluation time by developing Python-based data analysis and visualization scripts for tensile and pressure tests - Enabled FE simulations, measured by geometric accuracy through optical digitization and CAD modeling (CATIA) of test components - Improved EBSD image analysis, measured by reproducibility by programming MATLAB image-processing routines

- Improved medical image segmentation quality measured by consistency across datasets by creating lung segmentations from patient CT scans using MITK and ITK-Snap - Supported AI model refinement by generating validated segmentation ground truth for AI-based lung models in a clinical research context

- Ensured functional safety and reliability of hemodynamic medical devices, measured by successful release and re-testing rates through structured manual and automated testing - Verified, analyzed, and documented device defects, measured by audit compliance in accordance with SOPs and regulatory requirements - Validated new medical devices, measured by test coverage and approval readiness by executing predefined validation protocols and measurement procedures - Improved test consistency, measured by reduced variance through systematic execution of electrical, mechanical, and functional measurements - Maintained and updated device software, measured by stable operation post-deployment including controlled software upgrades and updates

Semester Thesis: Development of a Sterile End Effector for Manipulating Laparoscopic Instruments Developed and validated a sterile end-effector prototype for a Robotic Scrub Nurse concept (SASHA-OR), enabling safe handling of laparoscopic instruments: • Requirements engineering and concept selection using morphological analysis • Sterility-by-design: modular architecture and sterile foil barrier with electromagnetic coupling across the barrier • Prototyping & manufacturing: FDM (PC/ABS) + metal AM (DMLS) for the gripper body; iterative design for stability • Experimental validation: load, grasping and robustness tests; reliable pick-up/hand-over for different instrument diameters

Exchange semester at Stellenbosch University with a strong engineering-for-impact focus: Supported learners at a primary school through STEM tutoring (math, physics, basic technical problem-solving). Contributed to early-stage solar/electrification concept & prototyping efforts to help provide electricity in underserved communities. Actively promoted the intercultural components like collaborating across diverse backgrounds and strengthening clear, respectful communication in international teams. Strengthened skills in teamwork, leadership, and community-oriented problem solving in an international environment.

Bachelor's Thesis: Development of an intelligent intraoperative cleaning station for laparoscopic instruments: - Designed and built a working prototype of an intraoperative cleaning station intended for integration into an autonomous OR assistance system (SASHA-OR). - Mechatronics/Mechanical design: engineered a combined rotational + translational mechanism (servo motor + lead screw/linear guide) with a modular, service-friendly architecture. - Sensing & embedded control: implemented start/trigger detection using an inductive proximity sensor and Arduino Uno control logic. - Robotics software: implemented the cleaning routine for a Franka Emika Panda using ROS with script-based execution (Python3). - Validation: planned and executed systematic tests (setup & durability) across realistic contamination scenarios; achieved good total cleaning time depending on scenario. - Rapid prototyping: iterated using FDM and SLA 3D printing.