Ramp’s Alexander Shevchenko explains why his team conducts applied research while general-purpose models continue to improve. He describes a post-trained finance question-answering modelFine-tuning adapts a pretrained model by updating its parameters on examples chosen for a particular task or subject. used alongside a frontier model, reporting that it exceeded an earlier frontier model on a specific task while reducing latency.
He also discusses precomputed context cartridges and an agent swarmA multi-agent system contains multiple AI agents that interact, coordinate, divide work, or influence one another while pursuing tasks. used to find and reproduce security vulnerabilitiesA software vulnerability is a weakness in software that can permit an unintended or unauthorized effect on a system or its information. Its practical impact depends on the conditions under which it can be triggered.. Results and comparisons are presented as Ramp’s reported experience, rather than independently verified findings.
Watch on YouTube
