Its entire contents were two lines:
# hello-world My first repository.
My GitHub bio holds to this day. I still “can't code”, but I can build complex systems — data analysis pipelines, statistical simulations, ML & GenAI automations, and even web & Android apps.
This is just an example of what's possible in the current unprecedented era of AI-assisted coding, when combined with passion and domain expertise.
hello-world has been reborn as hello, brave new world: a single, self-contained
web page that boots into an interactive observatory. The words hello, world collapse
and detonate into an expanding supernova remnant, and from there you can pilot four live
instruments. Each one is not a mock-up but a real algorithm from my actual research —
the same physics and machine learning I work on, running in your browser.
→ Open the live experience (best on a desktop, in a dark room).
| Instrument | What it actually is | From the work | |
|---|---|---|---|
| ◉ | Supernova | A particle remnant whose shock shell expands as r ∝ t²ᐟ⁵ (Sedov–Taylor), false-coloured by plasma temperature. Click to detonate more. | X-ray evolution of supernova remnants — my MSc thesis on the LMC remnant SNR J0500−6512. |
| ◉ | Ising | A live 2D Ising model run by the Metropolis Markov-chain Monte Carlo algorithm. Drag the temperature across Tᴄ ≈ 2.269 and watch magnetisation emerge from chaos — a second-order phase transition, in real time. | My MSc Advanced Computational Physics project. |
| ◉ | ISM Scan | A vision-style detector sweeps a survey field and boxes bubble-like structures — supernova remnants, HII regions, wind bubbles, planetary nebulae — with confidences. | A homage to my vision–language model pipeline for detecting ISM structures (in prep). |
| ◉ | Network | An agent-based model: persona nodes spread misinformation across social ties while auditor nodes fact-check and recover their neighbours. Tip the network yourself. | My paper Simulating Misinformation Propagation in Social Networks using LLMs — Outstanding Paper, LASS @ ACM CIKM 2025. |
Switch instruments with the dock or keys 1–4. Click the field to interact. Hit ◍ whoami for the story.
I'm Vaibhav Shukla, an early-career researcher working at the seam of high-energy astrophysics and machine learning, based in Erlangen, Germany.
My MSc (FAU Erlangen–Nürnberg) was carried out at the Dr. Karl Remeis Observatory (ECAP) under Prof. Dr. Manami Sasaki, studying the X-ray evolution of supernova remnants in the Large Magellanic Cloud — combining single-object XMM-Newton spectral analysis with population-level statistics over the complete LMC sample. Alongside astrophysics I build ML and generative-AI systems: vision–language models for astronomical surveys, and multi-agent LLM simulations of information dynamics.
Selected work
- Simulating Misinformation Propagation in Social Networks using LLMs — LASS @ CIKM 2025 · Outstanding Paper · arXiv:2511.10384
- An Empirical Exploration of Socioeconomic Linguistic Bias in Text-to-Image Generation — SRW @ EACL 2026 · paper
- Vision–Language Model Detection of Interstellar-Medium Structures — in preparation (A&A / ADASS XXXVI)
- X-ray Evolution of Supernova Remnants in the LMC — MSc thesis (PDF)
Find me: academic site · all repositories · LinkedIn
It's one HTML file with zero dependencies — no build step, no framework, no tracking.
git clone https://github.com/whyvav/hello-world.git
cd hello-world
# just open it:
python3 -m http.server 8000 # then visit http://localhost:8000Or simply open index.html in any modern browser.
How it's built: a single <canvas> engine in vanilla JavaScript —
one animation loop, a shared starfield/nebula renderer, and four self-contained scenes.
The Metropolis updates, the Sedov–Taylor shock, the agent dynamics and the force-directed
graph are all computed live, ~60 fps, on your machine. The header banner is a single
animated SVG.