Glau Santana — Engineering leadership · Software architecture · AI
Engineering with purpose.
I build and lead engineering organizations in fintech and payments — and I research what happens to software architecture when machines write the code.
About
Glau Santana is a Head of Engineering with fifteen years across fintech, payments, marketplaces, and adtech. She leads cross-functional squads building large-scale loyalty and payments platforms, and is a doctoral researcher at the Polytechnic School of the University of São Paulo (EPUSP).
Her research adapts the Architecture Tradeoff Analysis Method (ATAM) for software generated by large language models — asking how we keep architectural intent intact when code is written by machines. Her work sits where engineering leadership, software architecture, and AI meet, and it carries one conviction: engineering with purpose.
Writing
Por que respeito explica mais resultados do que qualquer framework: segurança psicológica, feedback, papéis claros, limites e carreira no ritmo da vida.
Ler em português →Why respect explains more results than any framework: psychological safety, feedback, clear roles, boundaries, and careers that fit each season of life.
Read in English →Uma reflexão sobre força feminina, jornada dupla, autossabotagem, valorização e a transformação da presença em voz e poder de decisão.
Ler em português →A reflection on the double shift, self-sabotage, recognition, and the need to turn presence into voice and decision-making power.
Read in English →Frameworks não compensam fronteiras de time mal desenhadas. Team Topologies, Lei de Conway e a arquitetura organizacional por trás da agilidade real.
Read on Medium →Our models write code that compiles, passes its tests, and ships — and quietly erode the architecture underneath. Why we're measuring the wrong thing, and how to make architectural drift a number.
Research essay · ATAM-LLMResearch
My doctoral work adapts ATAM so that architecture steers code generation rather than being discovered after the fact. Architectural constraints are elicited, translated into prompt instructions before a model writes code, and then verified through static analysis and production telemetry.
The measurable distance between the architecture you decided on — Intent — and what production actually does — Reality.
The thesis: squads that generate code under architectural constraints produce measurably lower drift than squads working from free-form prompts.
A Roadmap for Identifying and Treating Non-Functional Requirements in Scrum Projects Through ATAM Practices
MSc in Software Engineering · Instituto de Pesquisas Tecnológicas (IPT), São Paulo
The dissertation proposed a roadmap that brings ATAM's architecture evaluation into the agile flow of Scrum, surfacing non-functional risks before code is written and monitoring them through delivery and production.
Mentorship
A mentorship program preparing women for leadership in engineering — across management, architecture, and software craft. Opening the path for the next generation of women who lead with purpose.
Contact
For speaking, research collaboration, advisory, or a conversation about engineering and AI:
LinkedIn · /in/glausantana
Email · glau@glausantana.com