Synthetic biology holds promise to engineer systems to treat diseases. One critical, yet underexplored, facet of designing such systems is the interplay between the system and the pathogen. Understanding this interplay may be critical to increasing efficacy and overcoming resistance against the system. Using the principles of synthetic biology, we engineer a strain of Escherichia coli to attract and intoxicate the nematode Caenorhabditis elegans. Our bacteria are engineered with a toxin module, which intoxicates the nematode upon ingestion, and an attraction module, which serves to attract and increase the feeding rate of the nematodes. When independently implemented, these modules successfully intoxicate and attract the worms, respectively. However, in combination, the efficacy of our bacteria is significantly reduced due to aversive associative learning in C. elegans. Guided by mathematical modeling, we dynamically regulate module induction to increase intoxication by circumventing learning. Our results detail the creation of a novel nematicidal bacterium that may have application against nematodes, unravel unique constraints on circuit dynamics that are governed by C. elegans physiology, and add to the growing list of design and implementation considerations associated with synthetic biology.