A successful application of a Machine Learning approach was implemented on the Uzen field dataset for the purpose of log quality control and reconstruction. In some of the wells Logs were significantly affected by the washouts, especially in the intervals of zones 12 and 13. ML based approach allowed to detect affected zones and correct the logs response. Fig.9 displays the log responses from the well, indicating the significant impact of borehole conditions (washouts) on RHOB, TNPH, and DTC data quality. Notably, the ML uncertainty flag correlates well with the caliper data, despite the caliper not being utilized as an input for the ML algorithm. This serves as evidence that the reconstruction was limited to the intervals of affected data. By employing the ML-based method, it was established that 15% of the wells encountered technical issues linked to the following factors in the data: • Incorrect count rate for neutron porosity computation • Inverted SP • Alfa-SP values outside of the limits. Overall, the application of the ML-based solution significantly reduces the time required for logs QC, enables the reconstruction of missing and affected data and reduces the influence of the human factor on the final outcome.