cv
This is a description of the page. You can modify it in '_pages/cv.md'. You can also change or remove the top pdf download button.
General Information
| Full Name | Huzaifa Arif |
| arifh@rpi.edu | |
| Phone | (518) 961-8482 |
Education
-
2021 - Ongoing Ph.D. in Electrical and Computer Systems Engineering
Rensselaer Polytechnic Institute, Troy, NY - {"GPA"=>"3.91/4.0"}
- Focus on Trustworthy AI, Privacy in Machine Learning, Federated Learning
- {"Advisor"=>"Dr. Alex Gittens"}
- Multiple first-author publications in top-tier conferences
- Research internships at IBM Research and Lawrence Livermore National Laboratory
-
2013 - 2017 B.S. in Electrical Engineering
Lahore University of Management Sciences, Lahore, Pakistan - {"GPA"=>"3.62/4.0"}
- Strong foundation in mathematics, signal processing, and computer science
Experience
-
Jun 2025 - Aug 2025 AI Research Extern - Trustworthy AI
IBM T.J Watson Research Center, Yorktown Heights, NY - {"Mentors"=>"Pin-Yu Chen, Ching-Yun Ko, Keerthiram Murugesan, Payel Das"}
- Developed parameter-efficient LLM safety alignment methods
- Investigated association leakage vulnerabilities in Large Language Models
-
May 2024 - Aug 2024 Data Science Intern
Lawrence Livermore National Laboratory, Livermore, CA - {"Mentors"=>"Bhavya Kailkhura, James Diffenderfer"}
- Developed evasion attacks for weather prediction models
- Published work accepted at AAAI 2025 workshop
-
Jun 2023 - Aug 2023 AI Research Extern - Trustworthy AI
IBM T.J Watson Research Center, Yorktown Heights, NY - {"Mentors"=>"Pin-Yu Chen, Keerthiram Murugesan, Payel Das"}
- Research on data leakage in residual neural networks
- Work accepted at SATML 2025 conference
-
Jun 2022 - Aug 2022 AI Research Extern - Trustworthy AI
IBM T.J Watson Research Center, Yorktown Heights, NY - {"Mentor"=>"Pin-Yu Chen"}
- Developed Reprogrammable-FL approach for federated learning
- Work accepted at SATML 2023 conference
-
Sep 2019 - Jan 2021 Research Assistant
Lahore University of Management Sciences (AdCommLab), Lahore, Pakistan - Developed Simulink model in MATLAB comparing 2x2 Alamouti vs DSTBC performance with BPSK/QPSK under Rayleigh fading
- Implemented and tested models on USRP N2310 and USRP X310 for over-the-air performance in indoor environments
- Achieved very low bit error rates (BER) through optimized signal processing techniques
- Developed Digital Auto Encoders to overcome DSTBC performance limitations compared to STBC
- Research focused on advanced wireless communication systems and space-time block coding
-
Aug 2018 - Jul 2019 Assistant Manager (Wireless Operations)
Pakistan Telecommunication Company Ltd, Lahore, Pakistan - Configured and maintained wireless network elements including microwave links, BSC, and NMS servers
- Extracted and analyzed alarm reports from BSC and MSC systems for improved network performance
- Optimized wireless network operations and troubleshooted connectivity issues
- Enhanced system reliability and operational efficiency across telecommunication infrastructure
- Managed technical teams and coordinated maintenance activities
-
Jul 2018 - Nov 2018 Project Engineer (E/I Engineering Intern)
Hascol Terminals Limited, Karachi, Sindh, Pakistan - Supervised technical operations in petroleum terminal engineering projects
- Managed electrical and instrumentation (E/I) systems for terminal operations
- Coordinated project implementation and technical oversight
- Applied engineering principles to optimize terminal efficiency and safety protocols
- Gained hands-on experience in industrial automation and control systems
-
Aug 2018 - Jul 2019 Research Intern
National Centre for Robotics and Automation (NCRA), LUMS, Lahore, Pakistan - Developed multi-camera tracking and mapping systems for UAV applications in canal environments
- Implemented computer vision algorithms for autonomous navigation and environmental mapping
- Worked on drone-based surveillance and monitoring systems for waterway infrastructure
- Applied robotics and automation principles to real-world environmental challenges
- Gained experience in UAV control systems and sensor integration
Publications
- DS FedProxGrad - Asymptotic Stationarity Without Noise Floor in Fair Federated Learning (arXiv 2025)
- Patching LLM Like Software - A Lightweight Method for Improving Safety Policy in Large Language Models (arXiv 2025)
- Reprogrammable-FL - Improving Utility-Privacy Tradeoff in Federated Learning via Model Reprogramming (SATML 2023)
- PEEL the Layers and Find Yourself - Revisiting Inference-time Data Leakage for Residual Neural Networks (SATML 2025)
- Group Fair Federated Learning via Stochastic Kernel Regularization (TMLR 2025)
- Forecasting Fails - Unveiling Evasion Attacks in Weather Prediction Models (AAAI Workshop 2025)
- DP-Compressed VFL is secure for Model Inversion Attacks (Preprint)
Patents
- Differentially Private Federated Learning using Model Reprogramming (US20240256894A1)
- Method for Quantifying Private Leakage in Pretrained Neural Networks (Pending)
- Patching LLM - Steering Large Language Models with Soft Prompting (Pending)
Honors and Awards
-
2025 - Founders Award of Excellence (Top 1% Graduate Students at RPI)
- 2nd Place at 3 Minute Thesis Competition
-
2024 - Belsky Award for Computational Science and Engineering (Top 6 Graduate Students)
-
2022 - PhD Qualification Exam (Rank Top-5)
Professional Service
- Reviewer for ICLR 2025
- Reviewer for ICASSP 2025, 2023
- Reviewer for AISTATS 2023
- Reviewer for IEEE MLSP 2023
Research Interests
-
Trustworthy AI
- LLM Safety and Alignment
- Privacy-Preserving Machine Learning
- Bias and Fairness in AI
-
Federated Learning
- Utility-Privacy Tradeoffs
- Model Reprogramming
- Differential Privacy