Ben
Consterdine

Working across Pendo's AI products: Novus, Leo (Pendo's in-product agent) and MCP tools.

Designed and built a digital twin of a modular indoor farm for terrestrial and orbital use, simulating crop growth and gas exchange at hourly resolution to plan crop mixes and assess viability.

Built a Bayesian network modelling app used by a consultancy to forecast project success. This included belief-propagation inference adapted from academic literature, and technical features such as import/export supporting legacy formats.

Primarily frontend work on Capsule CRM's EmberJS app. This included work across annual billing, comments, internationalisation, and an improved onboarding flow.