Recent advances in the treatment of the peptide design problem have led to the ability to select novel sequences given the structure of a peptide backbone. Despite these breakthroughs, issues related to the stability and functionality of these designed peptides remain sources of frustration. In this work, a novel method that addresses these issues for the computational design of peptides and proteins is introduced. The method is based on (i) in silico sequence selection for a template fold using a novel integer linear program (ILP) formulation and (ii) validation of fold stability and specificity through rigorous calculations of ensemble probabilities for the selected sequences. Through experimentally based functional analysis, the approach is shown to provide several peptide sequences with 6-7-fold improvement in activity over the synthetic therapeutic peptide compstatin, a 13-residue cyclic peptide that binds to complement component C3 and inhibits complement activation.