A method for implementing a speaker-independent speech recognition system with reduced latency is provided. The method includes capturing voice data at a carry-on-device from a user during a pre-flight check-in performed by the user for an upcoming flight; extracting features associated with the user from the captured voice data at the carry-on-device; uplinking the extracted features to the speaker-independent speech recognition system onboard the aircraft; and adapting the extracted features with an acoustic feature model of the speaker-independent speech recognition system.