Aryaman Sinha
Software engineer building scalable data platforms and ML systems
Gwalior, India
Joined April 2026
Summary
Published researcher in adversarial machine learning who develops and releases implementations of defense methods. ieee+2
Software engineer focused on building scalable backend and data-platform services, applying ML/GenAI techniques in production, and improving system observability. google+2
Maintains open-source repositories that reproduce and accompany academic work, enabling reproducibility and practical use of published defenses. github+1
Academic background from IIT Bhubaneswar with ongoing engagement in the research community including peer-review activity. google+2
Work
Education
Projects
Writing
Adversarial Defense with Local Robust Principal Component Analysis and Wavelet Denoising
January 1, 2023Introduces a localized RPCA-based reconstruction approach combined with wavelet denoising to remove adversarial perturbations, with experimental evaluation against state-of-the-art attacks.
NOMARO: Defending Against Adversarial Attacks by NOMA-Inspired Reconstruction Operation
January 1, 2022Proposes a communication-theory-inspired defense (NOMARO) that superposes an adversarially perturbed copy of an input with the original using a power allocation strategy to mitigate adversarial attacks on DNNs; evaluated against C&W and Square attacks.