Financial Services & Fintech · One Platform, One Profile, AI-Powered Execution
One Platform, One Profile, AI-Powered Execution
Delivering consolidation primitives while adding capacity to AI and ML workstreams
The Client · A leading peer-to-peer payments platform

Overview
A leading peer-to-peer payments platform engaged Taller to deliver consolidation building blocks and contribute engineering capacity to its AI and ML workstreams. The engagement was part of Taller’s multi-year, multi-pod partnership with the global payments company.
The Problem
The company set out to consolidate several separate payment platforms onto one shared foundation, and at the same time to run an AI program aimed at speeding up its own engineering using AI to generate tests, accelerate a major code migration, and shorten compliance investigations. Both efforts needed the same things: core building blocks delivered against the new consolidated architecture, and AI/ML engineering capacity to support the wider program. The client engaged Taller to help supply both.
The Solution
Taller delivered two of the consolidation’s core building blocks. The first was a post-purchase notifications component that plugged into the company’s AI-powered cashback and smart-receipts products, and involved migrating legacy C++ code to Java on the company’s internal framework. The second was a unified onboarding platform that collapsed more than thirty separate sign-up flows into a single, adaptable, localizable experience. On the AI side, Taller contributed engineers to the company’s Agent Builder (its self-service platform for creating AI agents) and provided machine-learning engineers supporting its large-scale ML systems for recommendations, ad ranking, personalization, search, and language processing.
The Impact
The program produced 10–30% reductions in coding effort, moved the checkout app from weekly to daily releases, cut AI-generated unit tests from 2–3 sprints to minutes, reduced a major code migration from 30 hours to 5, and shortened compliance investigations from days to minutes.
reductions in coding effort


