Anel Music 🍉

AI Engineering

LLM Post-training, Evaluation and RAG System Engineering My work encompasses scope definition, benchmarking, guardrail implementation, model fine-tuning, preference alignment, and model integration.

I played a key role in developing robust algorithms for data processing and sensor fusion, which were essential in tackling the challenges of robot localization. A significant achievement in this project was optimizing data validation procedures. By applying advanced similarity classification techniques, I was able to cut the processing time from about 8 hours down to just 15 minutes. Some tasks were more routine but just as important. These included unit and integration testing, code refactoring and code reviews.

I significantly contributed to the development of a smart parking facility by creating a system that accurately measures the dimensions of incoming vehicles. This was accomplished through the effective implementation of Semantic Segmentation and Photogrammetry techniques. The feedback resulted in a promotion just 6 months after joining the company.

I conducted research on multi-sensor fusion before designing a unique calibration target to accommodate various sensors, including cameras, lidars, and radars. Following this, I developed algorithms to extract distinctive features and determined the relative 6D pose of these sensors using advanced factor graph optimization techniques. This work was done alongside writing my thesis. The most rewarding part of the experience, however, was gaining insights from a community of exceptionally talented professionals.

Having read numerous books on startups and guides on commencing a company, my curiosity compelled me to further explore the factors that contribute to startups being profitable investments. During my time at fluxunit, I acquired valuable insights, discovering that a harmonious combination of multiple elements is crucial, with timing and the appropriate delegation of tasks to capable individuals being the most pivotal factors.

I had the wonderful opportunity to guide and mentor bachelor's students in robotics and programming. From conceptualizing robot behaviors to translating those ideas into functional code, my responsibilities encompassed a wide spectrum. By providing hands-on guidance and continuous support, I empowered these students to transform their theoretical knowledge into tangible robotics applications.

I was responsible for the verification and documentation of maintenance work. Once I understood the workflow I wrote a script that reduced my efforts from 16 hours to roughly 1 hour per week. This was the trigger to look for new challenges.

During my role as a student tutor in the university automation lab, I got the opportunity to contribute to research and practical applications in the field of robotics. I actively engaged with students, sharing my expertise and guiding them through complex projects. As a result, I played an instrumental role in cultivating a vibrant learning environment, fostering interest and maybe even passion for robotics among my peers.

Abschlussnote: 1.3

At M-Net, I worked on various programming assignments, documenting codebases and adding new functionalities. Ensuring up-to-date firmware, I conducted research, testing, and integrations. Additionally, I designed scripts that automated manual tasks, improving efficiency across workflows

During my internship, I had the opportunity to work on various software development tasks. Among these responsibilities, I took on the challenge of creating a process visualization for data center cooling systems. Through a strong passion for coding, I successfully developed a user-friendly and efficient solution. My efforts were greatly appreciated by the team, and the experience further solidified my commitment to pursuing a career in software development.