Piyush Pant

Piyush Pant

AI Researcher | Developer | Martial Artist

Saarbrücken, GermanyUniversität des SaarlandesCISPA Helmholtz Centerpiyushpant15@gmail.comPiyushWithPant | GitHubLinkedIn ProfileGoogle ScholarORCID iDWhatsapp

Who I Am

I come from a small Himalayan village near the Panchachuli peaks in the Kumaon region of Uttarakhand, India . My journey is defined by a leap of faith, leaving my roots to seek new challenges and expand my horizons in the pursuit of professional excellence.

My Journey

Master's Thesis

CISPA, Germany

Research on Stealthy Emotion-Targeted Subpopulation Attacks in RLHF, supervised by Prof. Dr. Xiao Zhang.

Research Assistant

CISPA, Germany

Deep diving into Adversarial ML and LLM security under Prof. Dr. Xiao Zhang.

M.Sc. Computer Science

Saarland University, Germany

Specializing in Artificial Intelligence and Machine Learning, with research focused on RLHF and robustness in large language models.

Research Assistant

School of Computer Science & Engineering, Sandip University, India

First research exposure in the field of AI & ML, worked on both AI & software development projects.

B.Sc. Computer Science

Sandip University, India

Completed my Bachelor's in Computer Science with specialization in Artificial Intelligence and Machine Learning, laying the foundation for my research journey.

School Life

K.V.S., India

Honored to have received my education at Kendriya Vidyalaya Sangathan, a prestigious institution in India where I completed my schooling.

The Himalayan Roots

Uttarakhand, India

Born and raised near the peaks, Lived almost in half of India

What I Work On

I am deeply passionate about advancing AI in a safe, ethical, and impactful direction. My research interests lie at the intersection of language, learning, and robustness. Lately, I've been drawn to the challenges of building Trustworthy AI and Adversarial Machine Learning, especially in the context of Large Language Models and Reinforcement Learning.

Exploring the intersection of NLP, RL, and Computer Vision
Building safe, reliable, and trustworthy AI systems
Studying adversarial attacks and defenses in language models
Improving model alignment, robustness, and generalization

"I aspire to pursue a PhD and contribute to meaningful research in AI safety and alignment. I enjoy working in collaborative, research-driven environments."

Beyond the Code

I love Martial Arts because of Goku ❤️. I have trained in various arts of combat in the last few years at my University and Clubs like Judo, BJJ, Boxing, Kickboxing and Wrestling. Currently, I am focusing on my Harai Goshi 🥋 :-)

I often find myself dwelling in the domain of Spirituality. I enjoy reading and discussing about the ancient wisdom of the Vedas, Yoga, and Bhagavad Gita.

"Even the gods desire to dominate others. Then why should I stop Shivji?

Because a true warrior doesn’t seek to dominate, but to discipline his own spirit. He knows nothing about surpassing others, only how to outdo himself. As he knows that - A strong man stands above and conquers all."