In my talk, I present my approach to solving the Ariel data challenge 2024. I employ a polynomial fitting to identify the ingress and egress's optimal start and end points. Then, building on methods proposed by Sergei Fironov (6th place winner), I employ 1D polynomial fitting (in time) and 2D polynomial fitting (in time and wavelength) to calculate the atmospheric spectra. I estimate the error in my predictions by relying primarily on the correlation between the mean error per planet and the standard deviation of the predicted atmospheric spectra. I calculate the atmospheric spectra separately for the ingress and the egress parts, giving more weight to the ingress calculation. I scale the predictions by multiplying them by a constant factor. I also smooth the predictions to various degrees depending on the standard deviation of the predictions per planet. Through an ensemble of one 1D polynomial fitting model and two 2D polynomial fitting models, I achieved a final score of 0.714 and the 4th place in the competition.