, Community Led Stage
Visdom is FOSSASIA's open source dashboard for watching machine learning experiments live, with over 10,000 GitHub stars. Over five months we rebuilt a large part of it in five stages, with 92 pull requests merged upstream
Storage abstraction layer: a DataStore interface so every environment save, load, delete and undo goes through one backend instead of raw JSON files.
Search layer: an experiment metadata store, an injection-safe query parser, search and field-by-field comparison.
Hyper-parameter layer: a dashboard with a run table, scatter matrix, parallel coordinates, a compare view and live updates.
Pytest: a suite of nearly 1500+ tests that uncovered real crashes and security bugs.
asyncio: disk I/O moved off the event loop and a new AsyncVisdom client, tackling an issue open since 2020 and reaching 2.1x the old client's throughput.
I'll show how each stage made the next one possible, the bugs we found along the way, and why small, reviewable PRs that never break the existing API are what got a mature project to accept this much change.
I'm Manik Khajuria, a Google Summer of Code 2026 contributor with FOSSASIA and a student at NIT Jalandhar (class of 2028). Since March 2026, I've had 92 pull requests merged into Visdom, FOSSASIA's open source dashboard for machine learning experiments with over 10,000 GitHub stars.
My work covers five layers of the project: a storage abstraction layer, an experiment search layer with an injection-safe query parser, a hyper-parameter tracking dashboard, a pytest suite of nearly 1500+ tests, and the migration of Visdom's server and client to asyncio. Along the way I also shipped security fixes, new ML plots such as ROC curves and confusion matrices, and the React 19 upgrade.
Outside open source, I'm a competitive programmer with over 2,000+ DSA problems solved. I've completed two internships, and I build with Python, React and Next.js.