Abstract: Antibody is becoming one of the most prominent molecular classes in both therapeutics and diagnostics. As the importance of protein therapeutics and diagnostics expands due to the rapid development of proteomics and genomics in medical applications, antibody engineering technologies leading to rapid antibody discoveries and to optimal physicochemical properties of the antibody molecules will hold a strategically prominent position in the biotechnology industry. We aim to build a superior antibody engineering platform overtaking the current state-of-the-art methodologies and to develop new antibody applications made possible only by the new antibody engineering development. Current antibody discovery relies on animal immunization-hydridoma or phage/yeast display technologies, which can now frequently uncover useful antibody molecules of human origin. But the shortcoming of the current methodology is the reliance of nature antibody sequence repertoires as sources for the lead antibodies, and thus the antibodies discovered are limited by the uncontrollable animal immune systems. One of the key developments in our lab is to construct protein recognition databases from the already vast collection of protein structures in the public domain. Machine learning and informatics algorithms built on the basis of the databases organize the knowledge extracted from the information and provide sequence designs that are more likely to be functional as antibodies. The information of successes as well as failures in sequence designs will form an integral part in the database and eventually feed back to the next round of computational designs. This cumulative enrichment in computer-organized knowledge will eventually result in a rationale-based methodology in synthetic antibody design previously unseen.