Fatih Ilhan

Bio

I am a Machine Learning Researcher at Apple. I earned my Ph.D. in Computer Science from Georgia Institute of Technology, where I was advised by Prof. Ling Liu. During my doctoral studies, I spent my summers as a Research Scientist Intern with the Hybrid Cloud Systems Research Group at IBM Research. Prior to Georgia Tech, I completed both my B.Sc. and M.Sc. in Electrical and Electronics Engineering at Bilkent University in 2019 and 2021, respectively.

Research: My current research focus is on-device machine learning optimization techniques and mobile-cloud hybrid computing for scalable application. I have also worked and collaborated on projects related to distillation for LLMs, AI safety, multi-agent AI, distributed learning, reinforcement learning and time series prediction/anomaly detection.

Teaching: I have served as the head TA of OMS CS6675 (Advanced Internet Systems and Applications) for four semesters and was selected as the Georgia Tech Head TA of the Year in 2025.

Please refer to my CV for the full record of my work and experience.

## Selected Publications

H3 Fusion: Helpful, Harmless, Honest Fusion of Pretrained-LLMs
S. F. Tekin, F. Ilhan, S. Hu, T. Huang, Y. Xu, Z. Yahn, and L. Liu. European Chapter of the Association for Computational Linguistics , 2026. (EACL) [paper] [code]

FedHFT: Efficient Federated Finetuning with Heterogeneous Edge Clients
F. Ilhan, S. F. Tekin, T. Huang, G. Liu, R. Kompella, G. Eisenhauer, Y. C. Lin, C. Pu and L. Liu. IEEE Conference on Cognitive Machine Intelligence , 2025. (IEEE CogMI) [paper] [code] [video]

Booster: Tackling Harmful Fine-tuning for Large Language Models via Attenuating Harmful Perturbation
T. Huang, S. Hu, F. Ilhan, S. F. Tekin, and L. Liu. International Conference on Learning Representations, 2025. (ICLR oral) [paper] [code]

Adversarial Attention Perturbations for Large Object Detection Transformers
Z. Yahn, S. F. Tekin, F. Ilhan, S. Hu, T. Huang, Y. Xu, M. Loper, and L. Liu. International Conference on Computer Vision , 2025. (ICCV) [paper] [code]

Resource-Efficient Transformer Pruning for Finetuning of Large Models
F. Ilhan, G. Su, S. F. Tekin, T. Huang, S. Hu, and L. Liu. IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2024. (CVPR) [paper] [code] [video]

Adaptive Deep Neural Network Inference Optimization with EENet
F. Ilhan, KH. Chow, S. Hu, T. Huang, S. F. Tekin, W. Wei, Y. Wu, M. Lee, R. Kompella, H. Latapie, G. Liu, L. Liu. IEEE/CVF Winter Conference on Applications of Computer Vision, 2024. (WACV) [paper] [code]

Lazy Safety Alignment for Large Language Models against Harmful Fine-tuning
T. Huang, S. Hu, F. Ilhan, S. F. Tekin and L. Liu. Thirty-seventh Conference on Neural Information Processing Systems, 2024. (NeurIPS) [paper] [code]

LLM-TOPLA: Efficient LLM Ensemble by Maximising Diversity
S. F. Tekin, F. Ilhan, T. Huang, S. Hu and L. Liu. ACL Conference on Empirical Methods in Natural Language Processing, 2024. (EMNLP findings) [paper] [code]

ScaleFL: Resource-Adaptive Federated Learning with Heterogeneous Clients
F. Ilhan, G. Su and L. Liu. IEEE/CVF Conference on Computer Vision and Pattern Recognition, 2023. (CVPR) [paper] [code] [video]

Lockdown: Backdoor Defense for Federated Learning with Isolated Subspace Training
T. Huang, S. Hu, F. Ilhan, S. F. Tekin and L. Liu. Thirty-seventh Conference on Neural Information Processing Systems, 2023. (NeurIPS) [paper] [code]

Markovian RNN: An Adaptive Time Series Prediction Network with HMM-based Switching for Nonstationary Environments
F. Ilhan, O. Karaahmetoglu, Ismail Balaban and S. S. Kozat. IEEE Transactions on Neural Networks and Learning Systems, 2021. (IEEE TNNLS) [paper] [code]

Modeling of Spatio-Temporal Hawkes Processes with Randomized Kernels
F. Ilhan and S. S. Kozat. IEEE Transactions on Signal Processing, 2020. (IEEE TSP) [paper] [code]