Kargi Chauhan

About Kargi Chauhan

Not big on self-promotion. Here’s the TL;DR of me and my work - no hype, just facts.Every challenge is a step toward something transformative. I see myself in the equation lim(x → ∞) f(x)= ∞.From developing advanced ML models to pioneering solutions for NASA and creating immersive digital experiences, I’m driven by a vision to make technology powerful and purposeful. Each project has been a journey of innovation, from enhancing satellite safety systems to reimagining education in virtual environments.As an ML Engineer, I fine-tune language models and integrate cutting-edge optimizations to elevate real-world applications. My work is fueled by the belief that technology should bridge gaps, solve real problems, and create connections that matter.More about me at kargichauhan.github.io or reach out to me at k••••••••@gmail.com

Experience

1. Graduate Researcher

   Aiea Lab

   May 2025 — Present

   Trained verifiable RL agents (PPO/A2C) in synthetic environments using PyTorch + Gymnasium, optimizing network architectures and GAE parameters to achieve a +13.4 average reward improvement.• Built an autonomous driving simulation infrastructure on Nautilus GPU clusters, deploying CARLA with automated traffic generation, and developed a robust benchmarking harness for large-scale evaluation.• Researched formal verification techniques to detect sensor spoofing and improve robust perception pipelines in safety-critical autonomous driving scenarios.• Experimented with GAIA-2 for video generation to extend AV simulation capabilities.Publication: Proud to share that this work culminated in a paper accepted at NeurIPS 2025.

Education

• University of Arizona

  Bachelor of Science - BS, Information Science Emphasis: Machine Learning and Data Science

• University of California, Santa Cruz

  Master of Science - MS, Natural Language Processing

• CodePath

  Courses: TIP 102

• Navrachana International School, Vadodara

  (International Baccalaureate Diploma Program - IBDP)