Cyril Halin

• Since 01.2025 : Hands-on development of a custom RAG pipeline using LangChain and fine-tuned LLMs for personal experimentation in document summarization and QA systems. Implementation of a vector search with FAISS and integrated GPT-based models for natural language queries. • 10.2024 : Pursuing Google Cloud Platform (GCP) "Professional Machine Learning Engineer" certification. Certified Google Cloud Platform (GCP) "Associate Cloud Engineer" certification.

• Since 01.2024, AI/ML Software Engineer. Machine Learning and Deep Learning. 3D scene reconstruction. 2D and 3D object detection. Synthetic data generation. Developed and deployed GenAI models, diffusion model. Fine-tuning diverse AI model and deployment in cloud for production. Cloud infrastructure setup. Patent registered. Skills : Python, Computer vision, PyTorch, Transformers based models, NeRFs, Gaussian Splatting, GenAI, Azure Cloud • 1 year, C/C++ automotive embedded software developer and software responsible for a series project. Feature owner, software requirement analysis, Coding guidelines, Software architecture, Design Pattern, Continuous integration with Jenkins, Unit-test implementation and automation with Google-Test, Documentation, Reviews. Parallel programming and GPU-offloading. Management of the software tasks. Skills : C/C++, computer vision, Git, Google-test, Jenkins, Doors, Agile and scrum, Jira, Track and Release, CAN bus, automotive ethernet. • 2 years, Python, C/C++ robotics software developer in ROS. Software development for AI, data preparation on Azure and ROS. Training and adapting Yolo CNN for people detection. Continuous integration and continuous development responsible. Setup of CI/CD Tool-chain with Git, Jenkins, coding guidelines, Sonarqube static code analysis, release responsible. Skills : Python, C/C++, CNN, Tensorflow, Yolo algorithm, CI/CD pipeline, Git, Jenkins, Azure-Cloud, agile and scrum, Vector (Canoe), CAN bus, ROS, Sonarqube, Jira. • 4 years, C/C++ automotive embedded ECU software developer. Series and prototype agile project. Feature owner, software requirement analysis, Coding guidelines, Software architecture, Design Pattern, Continuous integration with Jenkins, Unit-test implementation and automation with Google-Test, Documentation, Reviews. Skills : C/C++, Computer vision, Git, Google-test, Jenkins, Helix QAC, Doors, Agile and scrum, Jira, ADTF, Vector (Canoe), CAN bus, Automotive Ethernet.

Master Thesis at Bosch in the Computer Vision Research Department Development of video based smoke detection algorithm. Research of state of the art algorithms for smoke detection. Development of new methods and new features for video based smoke detection with computer vision algorithms in C++ and Matlab

Matlab and C++ software development and simulation for ADAS with lidar as reference for a camera in an autonomous driving vehicle, processing and filtering of lidar data, State flow development for ADAS.