Armeet Singh Jatyani

CVPR 2025 paper (and arXiv preprint) proposing a unified, neural‑operator approach to compressed‑sensing MRI that is robust across undersampling patterns and image resolutions; shows improved SSIM/PSNR vs baseline and provides code.

Patent application (2022) for a head safety / helmet device — related to earlier design and outreach work described in TEDx talk.

Argues that current sandbox coding agents (Cursor, Codex, Copilot) fall short because they cannot run code or access full developer context; outlines desired features (multiplayer, better UI, true dev environments) and mentions an internal tool (Chopin).

Explains switching from fish to bash for speed and minimal startup overhead, and recommends a lightweight prompt (starship) for essential status info.

A reflective piece on productivity culture arguing against equating self‑worth with constant productivity and advocating for authenticity.

Describes a GPU‑accelerated 3D MRI volume renderer written in Rust compiled to WASM, with links to the GitHub repo and live demo.

Discusses why human-generated, authentic writing remains valuable even as AI‑generated content becomes widespread.

Argues the web provides a fast, broadly capable platform for many applications and highlights the benefits of WebAssembly and rapid development.

Practical guide and code for setting up Telegram notifications for machine‑learning training runs, including a PyTorch Lightning callback implementation.

Step‑by‑step guide for running multi‑node PyTorch training on TACC Vista with SLURM scripts, environment variable setup, and testing tips.