You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hybrid CNN-Transformer model for automated exoplanet transit detection on NASA Kepler light curves. Features dual scale CNN branches, Transformer based sequence modelling, Grad CAM + SHAP + attention explainability, and Monte Carlo Dropout uncertainty quantification. 5 fold cross validated with baseline comparisons.
Interactive site documenting my TIFR NIUS 2025 astrophysics research — CCD image calibration pipeline (bias frames, aperture photometry, image subtraction) and exoplanet transit detection using TESS light-curve data, PLD systematics correction, and Box Least Squares period finding.
An end-to-end AI-driven pipeline for automated exoplanet detection, using a hybrid framework of 1D CNNs and Box Least Squares (BLS) to filter, classify, and physically model transit signals from noisy, crowded-field TESS light curves.
Report on GJ 1214 b: a JWST MIRI phase-curve spectrum inverted to per-bin brightness temperature, with propagated uncertainties, compared against Kempton et al. (2023)'s own published values.
Report on HAT-P-11b: real CARMENES helium 10830 A transit spectroscopy, with our own band-averaged statistic reported alongside the published Allart et al. (2018) measurement.
Report on WASP-121 b: a JWST phase-resolved emission spectrum, with each brightness-temperature point's own posterior uncertainty propagated into the day-night contrast.
Report on K2-18 b: system parameters and a combined JWST NIRISS+NIRSpec+MIRI spectrum, running a simple band-vs-continuum test on the disputed CO2 claim.
Report on TRAPPIST-1 e: a flat-line test on its decontaminated JWST NIRSpec/PRISM transmission spectrum, reported alongside Espinoza et al. (2025)'s own atmosphere constraints.