Throughout my 10 years of experience in Machine Learning and 6 of focused Deep Learning research in both ASR/NLP domains, I have acquired an in-depth understanding of the intricacies involved in developing cutting-edge models. I am particularly drawn to the LLM aspect of the role, where I can contribute my expertise in high-impact research projects and ensure the delivery of publication-quality work. My comprehensive knowledge of the latest advancements in NLU, coupled with my ability to replicate state-of-the-art models from conference and journal publications, will enable me to contribute to Dimension's mission of pushing the boundaries of AI technology.
I possess a strong foundation in understanding the applications and implications of various neural network types, architectures, and loss mechanisms. I am proficient in popular LLM frameworks GPT, Claude, Haystack, Langchain and deep learning frameworks such as PyTorch, TensorFlow, and Keras. I have extensive coding experience in Python, C/C++ and am familiar with Flutter and Typescript. Additionally, I am adept at navigating UNIX-style systems and leveraging data-driven insights to iterate rapidly on model prototypes. While not mandatory, I have knowledge of edge computing and ASR which can be valuable in certain contexts. Furthermore, my published IEEE journals and NIPS/ICASSP papers in Deep Learning Research, particularly focusing on neural networks, demonstrate my commitment to advancing the field. I reached top-ranked positions at AI challenges such as Chime3, Chime4, DCASE, as well as IEEE echo-cancellation. This achievement highlights that I can deliver impact solutions and have an eye for the details.
I have had the privilege of working on diverse AI projects that have pushed the boundaries of AI technology. Notably, at GoodAI and Flapjack GmbH, I led the development of AI search and information retrieval using GPT architectures such as Claude, ChatGPT, LLama, Mistral and vectorized databases. This work demonstrated my ability to incorporate state-of-the-art technology into industrial products and implementing both privacy-aware and cloud GPT pipelines into production systems. During my activity as a technical lead at Robotic Eyes GmbH, I successfully designed and implemented a vision system for AI digitization in prefabrication factories (www.chekker.com). This system revolutionized the construction industry by enabling real-time mapping of digital plans onto construction sites. The successful launch of this system stands as a testament to my technical leadership and ability to deliver impact solutions, leading a team of more than 30 people.
I am not only passionate about technical excellence but also possess exceptional communication and collaboration skills. I excel at conveying complex technical concepts to both technical and non-technical stakeholders, fostering clear understanding and driving alignment. As a collaborative team member, I actively share knowledge and contribute to a positive team culture, recognizing the importance of diverse perspectives in achieving remarkable outcomes.