Summary
Sahil Verma is a research-focused machine learning engineer and soon-to-be PhD from the University of Washington with a decade of experience applying ML to safety, interpretability, and robustness challenges. He has held research internships at Microsoft and Amazon where he tackled multilingual jailbreak defenses and controlling LLM outputs amid conflicting context, and earlier work spans explainability for recommenders, fairness, and symbolic CSP solving. His academic and industry blend gives him a strong footing in both practical system constraints (TensorFlow shape issues, floating-point tuning) and conceptual ML safety problems. Based in Seattle, he consistently translates theoretical insights into deployable mitigations and tooling for real-world models. Notably, his recent Microsoft work contributed to a public arXiv paper on multilingual jailbreak defenses, reflecting an emphasis on cross-lingual robustness often overlooked in safety research.
10 years of coding experience
Doctor of Philosophy - PhD Computer Science, Doctor of Philosophy - PhD Computer Science at University of Washington
Indian Institute of Technology Kanpur
English, Hindi