Raptors exert top‐down influences on ecosystems via their effects on prey population dynamics and community composition. Most raptors are sympatric with other predators, thus complicating our understanding of their relative influence in these systems. Estimates of kill rates and prey biomass recycling have been used as predation metrics that allow quantitative comparison among species and assessment of the relative role of single species within complex food webs. Few studies have produced findings of kill rates or prey biomass recycling for raptors. We used a supervised machine learning algorithm to behaviourally classify high resolution accelerometer informed GPS locations of tagged adult non‐breeding martial eagles Polemaetus bellicosus in the Maasai Mara region of Kenya to estimate kill rates and prey biomass recycling. Eagle locations classified as feeding were clustered using distance and time thresholds to identify kills and calculate kill rates. Identified kill sites were quickly ground‐truthed to confirm kills and identify prey species. We estimated kill rates for martial eagles at 0.59 kills day‐1 for males and 0.38 kills day‐1 for females, and we estimated biomass recycling per ground‐truthed kill at 1796 g for males and 3860 g for females. From our sample of identified ground‐truthed kills, ‘gamebirds' was the most frequently recorded prey category for male eagles and ‘small ungulates' was the most frequently recorded prey category for female eagles. These results position martial eagles close to sympatric mammalian top predators in trophic pyramids and provide evidence for their classification as a top predator.