AparajeetS/CDC
Numerical solvers, support data, and reproducibility utilities for the Configurational Dispersal Cosmology Paper II draft.
Repositories
AparajeetS/Mineral-World-Model
3D mineral-system simulation and inference with geophysical data, neural inverse models, Bayesian inference, and spatial validation.
AparajeetS/Icarus
LaTeX/PDF exam practice engine with timed tests, grading, detailed reports, and source-linked study cards.
AparajeetS/aparajeets.github.io
Aparajeet Shadangi's personal research portfolio: independent research, numerical cosmology, and scientific computing.
AparajeetS/phason-labs
Phason Labs research website and controlled benchmarks for ML training diagnostics and evaluation reliability.
AparajeetS/phason-evidence-audits
Seven executable ML evidence audits: calibration, distribution shift, inference correctness, and software reliability, with reproducible artifacts.
AparajeetS/fsigma8-growth-consistency
Reproducible cosmological growth-rate benchmark with provenance, covariance accounting, MCMC/bootstrap diagnostics, and numerical validation.
AparajeetS/AparajeetS
AI safety, evaluation, and research engineering | Aparajeet Shadangi
AparajeetS/CDC-GeoK1-research
Public progress dossier for the CDC-GeoK1 geometric cosmology research program
AparajeetS/marginal-baseline-eval
Open protocol for testing whether ML and AI-safety metrics add information beyond cheap baselines.
AparajeetS/Traintools
Agent-ready PyTorch training diagnostics: gradient health, GNS, batch inspection, data quality, plasticity, and early stopping.
AparajeetS/LLMs-from-scratch
Implement a ChatGPT-like LLM in PyTorch from scratch, step by step