I am a PhD in Machine Learning specialized in identifying and addressing the limitations of modern ML models. Beyond my research, I enjoy distilling complex AI papers into actionable insights for the community on LinkedIn.
My research outcome
- Test-time defense through model reprogramming – First prompt-based defense technique of its kind. Link
- Multimodal robustness analysis – Demonstrated that latent diffusion models exhibit distinct vulnerability patterns compared to traditional perturbation-based attacks. Link
- Adversarial detection at scale – Created an open-source benchmark dataset (OpenOOD contribution). Link
Research interests: Model robustness • reasoning • deepfake • test-time learning • autonomous systems • foundation models • pattern recognition • large-scale AI • multimodality • self-improving architectures • AGI.
📫 How to reach me: My mail
- 🔭 I’m currently working on transformers and foundation models.
- 🌱 I’m currently learning agengtic AI and evolutionary AI.
- 👯 I’m looking to collaborate on ML research projects. (Collaborate with other AI-obsessed people and create something awsome.)
- 🤔 I’m looking for mock interview practice. Link
- ⚡ Fun fact: I won a tournament in kickboxing.