Daniel Panea Lichtig

Designed and deployed AI integrations and automations using Retrieval Augmented Generation (RAG) workflows tailored to client-specific use cases. Delivered privacy-centric solutions by implementing on-premise AI/ML models, ensuring full compliance with data protection requirements. Developed and fine-tuned machine learning models for classification, prediction, and pattern recognition tasks using frameworks such as PyTorch and Hugging Face. Created robust, custom Python algorithms to extract actionable insights from high-volume, noisy, and unstructured data sources.

Led R&D projects from concept and simulation to implementation and optimization. Developed & implemented custom computational models and algorithms, improving optical system performance (4× increase in measurement range, 5× accuracy boost). Automated simulation data extraction and analysis through custom data pipelines, enhancing efficiency. Advanced data analysis in Python (NumPy, SciPy, OpenCV, PyTorch, Pandas) to optimize optical systems. Proficient in version control (Git), GPU acceleration, and Autodesk Fusion 360.

Developed a high-performance GPU-accelerated algorithm for efficient image decomposition into a basis of Zernike polynomials using C++ and CUDA, improving computational speed and efficiency and enabling a fast and accurate analysis of wavefront measurements of silicon wafers for microchip manufacturing.