ML Researcher, Apple, United Kingdom.
Fazilet Gokbudak is a machine learning researcher at Apple, specializing in generative models, computational photography, and inverse rendering. She received her PhD in Computer Science from the University of Cambridge in 2024, where she worked on neural rendering and efficient image manipulation techniques. Her academic path began at Bogazici University with a high honors degree in Electrical and Electronics Engineering, followed by an MSc with distinction from the University of Edinburgh. Her industrial journey includes pivotal contributions at Amazon as an Applied Scientist Intern. Fazilet actively promotes diversity in AI as a co-chair at Women@CL and reviewer for prestigious AI Awards. Her impactful work has been recognized through awards, publications in top-tier Awards like ECCV and SIGGRAPH, and cutting-edge research contributing to sustainable visual technologies.
Author’s Profile
Education
Fazilet Gokbudak pursued her undergraduate studies in Electrical and Electronics Engineering at Bogazici University (2014–2018), graduating with High Honors and multiple scholarships. She then completed her MSc in Signal Processing and Communications at the University of Edinburgh in 2019, earning a distinction. Between 2020 and 2024, she conducted her PhD research in Computer Science at the University of Cambridge under full funding from the Computer Laboratory Studentship. Her doctoral work focused on generative neural techniques for photorealistic appearance manipulation and inverse rendering. Fazilet’s academic training spanned signal processing, computer vision, and machine learning. Throughout her education, she demonstrated a consistent commitment to innovation, excellence, and sustainability in computational research, reflected by her active participation in research projects and her growing publication record.
Experience
Fazilet currently serves as an ML Researcher at Apple (since October 2024), where she develops advanced camera algorithms to enhance mobile photography. Previously, she interned at Amazon (July–November 2022), working on high-fidelity conditional image generation using GANs with local histogram losses for skin tone realism. She also worked as a part-time Research Assistant at the University of Cambridge (Nov 2020–Jan 2022), where she led the Cambridge team on a joint blackgrass detection project using deep convolutional networks. Her work achieved 80% classification accuracy on novel agricultural datasets, showcasing practical impacts of AI in sustainable agriculture. Her career blends foundational machine learning research with real-world, production-level deployments.
Awards and Honors
Fazilet has earned multiple accolades recognizing her academic excellence and research impact. At the University of Cambridge, she received the prestigious Computer Laboratory Studentship (2020–2024), which fully funded her doctoral studies. She also earned a Graduate Student Travel Award from Queens’ College in 2023 to attend SIGGRAPH, a top-tier graphics Award. At Bogazici University, she graduated with High Honors and received the Dean’s High Honor Certificate in 2018. She was also a recipient of the TEKFEN Holding Private Scholarship (2009–2018) for her outstanding performance in national exams, ranking 185 out of over two million candidates. Her high school career concluded as Valedictorian, earning Summa Cum Laude. These honors reflect her long-standing dedication to academic excellence and societal impact.
Research Focus
Fazilet’s research spans several cutting-edge areas in machine learning and computer vision, with a core focus on generative models, neural rendering, and appearance manipulation. Her PhD work centered on data-efficient methods for realistic image generation, including patch-based transformations and BRDF modeling. She explores inverse rendering techniques that enable physically consistent reconstructions of visual scenes, contributing to advancements in sustainable graphics systems. At Apple, her focus includes developing intelligent algorithms for mobile camera systems, aligning technological performance with energy efficiency and visual realism. Her research also supports sustainable tech by optimizing neural computations and minimizing training overhead, critical for eco-conscious AI applications. She brings a unique blend of academic depth and practical innovation.
Publication Top Notes
Multispectral Fine-Grained Classification of Blackgrass in Wheat and Barley Crops (2025, Computers and Electronics in Agriculture)
Co-authored with M. Darbyshire et al. This study develops a deep learning model using multispectral image data to accurately identify blackgrass weed species in cereal crops. Fazilet was responsible for model training and evaluation, contributing to a high-accuracy, fine-grained classification system. The work supports sustainable agriculture by enabling targeted herbicide use and reducing environmental impact.
Spatial Receptor Allocation for a Multiple Access Hub in Nanonetworks (2019, IEEE Transactions on Molecular, Biological and Multi-Scale Communications)
Fazilet’s early research explored theoretical models for receptor allocation in nanoscale communication networks. The study introduces simulation-based methods to optimize signal clarity and reduce cross-interference in molecular communication systems, marking a foundational step in nano-IoT frameworks.
Hypernetworks for Generalizable BRDF Representation (2024, European Conference on Computer Vision)
This paper presents a novel hypernetwork design that enables generalization across various materials when estimating Bidirectional Reflectance Distribution Functions (BRDFs). Fazilet contributed to network architecture and experimentation, demonstrating the model’s ability to capture complex material appearances using fewer parameters, facilitating efficient rendering in graphics pipelines.
Data-efficient Neural Appearance Manipulations (2025, in preparation)
This upcoming work proposes neural models for photo editing that require less training data and computational power. Fazilet introduces modular architectures that allow intuitive and efficient image edits, especially suited for low-resource devices. The approach balances realism and sustainability by minimizing hardware dependency.
Physically Based Neural BRDF (2024, arXiv preprint)
Fazilet co-authored this paper which merges physical reflectance models with neural networks to improve accuracy in appearance modeling. The technique enhances inverse rendering applications and supports high-fidelity visual reconstruction by incorporating physical consistency into the learning process.
One-shot Detail Retouching with Patch Space Neural Transformation Blending (2023, ACM SIGGRAPH CVMP)
The paper introduces a one-shot learning-based approach for image retouching that uses patch-space neural blending. Fazilet’s contributions include model design and testing, enabling fast, high-quality transformations from a single reference image—a technique ideal for low-data environments.
Patch Space Neural Field based Transformation Blending (2022, CoRR)
A precursor to her 2023 SIGGRAPH CVMP paper, this work investigates neural field-based image transformation using patch-level information. It showcases the potential for detail-preserving edits with minimal training samples, advancing low-data generative editing.
Conclusion
Fazilet Gokbudak is an exceptionally strong candidate for a Best Researcher Award in the fields of Machine Learning, Computer Vision, and Generative AI. Her multidisciplinary background, strong publication record, and industry-academic synergy position her as a next-generation leader in AI research.
