I create "down to the metal" software for science and engineering.
My career has centered around software the assists the expert in decision making, or aids the user in making unbiased decisions in complex problem spaces.
Many (most) problems in business and industry are defined by a decisison space that is too complex. No perfect solution is possible. With my skillset, I learn a domain and model software that assists the user in choices that are optimal, rather than "perfect"
Examples include:
Genetic Algorithms for choosing pharmaceutical candidates.
Custom algorithms for content scheduling for commercial and music rotation in the broadcast industry. Applicable to trucking, general scheduling and yield management.
Electronic Data Maintenance for Airline pilots in fullfilling pre/post landing FAA requirements.