Resilient payment platforms
Designing backend flows for payments, reporting, reconciliation, and failure isolation in distributed services.
Edoardo Fratus
I am Edoardo Fratus, a Software Engineer III in Milan working on high-scale fintech systems at Satispay. I like the part of engineering where product pressure, reliability, and clean system design all meet: payments, reconciliation, fan-out processing, circuit breakers, and the boring-but-critical details that keep platforms calm under load.
Recently I have been leading end-to-end delivery for major Welfare and Flexben features, coordinating Product, frontend, mobile, and backend work while staying hands-on in backend technical analysis and implementation. I also use AI in a concrete way: building agents and local workflows that remove repetitive engineering work without removing engineering judgement.
What I work on
Designing backend flows for payments, reporting, reconciliation, and failure isolation in distributed services.
Using event-driven fan-out, AWS services, and Virtual Threads to improve throughput in I/O-bound systems.
Building Claude subagents, local AI experiments, and workflow automation that solve real engineering bottlenecks.
Turning ambiguous product goals into technical analysis, coordinated delivery, launch readiness, and production follow-up.
Case studies
Problem: A major Welfare/Flexben feature needed to move from product discovery to production without losing alignment across multiple teams.
Approach: Owned backend technical analysis, translated product needs into implementation plans, and coordinated frontend, mobile, and backend integration.
Impact: Delivered through a six-month lifecycle with cleaner cross-team execution and a stronger bridge between product goals and backend reality.
Read case studyProblem: Recurring low-level and boilerplate engineering tasks were slowing teams down and consuming attention better spent on design quality.
Approach: Created Claude subagents and practical AI workflows for code generation, analysis, scaffolding, and repetitive task automation.
Impact: Moved AI from experimentation into concrete internal engineering workflows used beyond a single personal setup.
Read case studyProblem: Flexben year-end operations required simultaneous processing of many closing workflows under bursty load.
Approach: Designed an event-driven fan-out system and paired it with resilience patterns for safer high-volume execution.
Impact: Improved scalability and operational confidence for time-sensitive order closing flows.
Read case studySelected stack
How I work
I like taking a feature from ambiguous product need to technical plan, implementation, launch, and follow-up.
Leadership is most useful when it is grounded in design tradeoffs, code review, production constraints, and real delivery pressure.
LLMs, agents, and local AI experiments are useful when they reduce repetitive work, improve analysis, or make migrations safer.
Selected experience
Satispay
Mar 2026 - present
Leading end-to-end delivery of major Welfare and Flexben features, translating product needs into technical specifications and coordinating frontend, mobile, and backend teams.
Satispay
Feb 2025 - Mar 2026
Designed and implemented high-impact Welfare Flexben features, including core payment platform work, report reconciliation, fan-out order processing, circuit breakers, and virtual-thread adoption.
Satispay
2023 - 2025
Co-built the Satispay Fringe Benefits backend from scratch, focusing on Order and Purchase flows and supporting EUR 30M in transactions in the first three months.
Accenture
2023
Developed a Python microservice for document defect detection using computer vision, YOLO, and deep learning.
Writing
Contact