In the current digital landscape, enhancing authentication mechanisms is essential to secure user data and prevent unauthorized access. The present paper discusses a two-layer authentication protocol based on a combination of keystroke dynamics and one-time password (OTP) to enhance security. The proposed system employs the biometric typing characteristics with PiNet-BiLSTM to make out a confidence rating. This is more precise and safer than conventional password-based systems. It is highly unlikely that brute-force attacks will occur when the OTP verification procedure is time-sensitive. The two-factor authentication system ensures that when one layer is breached, the other layer will ensure that the security remains in place. A new activation function has been introduced in the deep learning design to handle non-linearity in a better way. This increases the accuracy of classification since it is possible to discriminate features better. Type-Safe Authenticator is used in the proposed system to secure the authentication information. This system relies on the encryption of Advanced Encryption Standard (AES), which makes the storage and transmission of sensitive data secure. The multi-level authentication system adopted in this research is more secure, reliable, and therefore an apt choice in the area of digital system safety.