Ariel Katz

Designed deep learning model (PyTorch) classifying complex mass spectrometry patterns, achieving human-expert accuracy with 100x faster throughput (Master's Thesis). Developed standardized Conda packaging workflow adopted across lab projects ensuring reproducibility. Built end-to-end OCR+LLM solution (Llama, Gemini) automating document extraction, correction, and summarization (Freelance). GitHub: github.com/Kezahle/parent_ion_classifier

- Developed distributed, high-concurrency, fault-tolerant block storage for high-performance, data-critical Linux systems - Implemented user-space port of core kernel components with comprehensive CI/CD testing, achieving 10x performance improvement while maintaining operational reliability

Led product strategy for a $150M/year Intent AI product, driving a 35% revenue increase and leading a multidisciplinary team of 12

- Led a project team focused on functionality and sensing of autonomous driving projects - Managed the product for the lane detection algorithmic team (27 engineers)

Delivered software projects for clients like McAfee, Texas Instruments and Alfatech, focusing on system expansion, network optimization and vulnerability research

Co-founded and led an AI-driven content distribution startup, overseeing product, tech and GTM strategy

Developed kernel-level cybersecurity products with a focus on Windows internals and real-time applications

Research Thesis - “Leveraging Deep Learning for Signal Classification in Data-Independent Mass Spectrometry” (in collaboration with the Weizmann Institute of Science); publication is in progress