Dangfeng develops methods that translate conservation decision problems into Markov Decision Process (MDP) models using Large Language Models and the RDDL formalism. His work aims to automate the pipeline from natural language problem descriptions to validated decision models, supporting rigorous optimisation under uncertainty. This is particularly important for conservation planning, adaptive management, and other high-stakes decision-making settings. His goal is to make advanced decision-support tools more accessible and easier to deploy in real-world environments.