About Me
Senior Solutions Architect working across Guidewire core systems and digital platforms — with a long-running research track in ML, RL, and agentic systems. I design delivery paths that teams can actually ship and operate.
Hi — I’m Axel Domingues. I’m a Senior Solutions Architect with over a decade of experience working on enterprise insurance platforms, digital experiences, and large-scale delivery programs.
Most of my work lives where programs either stabilize and scale — or quietly drift into chaos:
Most background is grounded in computer science and engineering (BSc + MSc), and most of my professional work has lived inside complex, high-change environments where reliability and execution discipline matter more than theoretical correctness.
It started with classical ML fundamentals, moved through deep learning and RL failure modes, and has recently focused on agents as production systems: evaluation, tool safety, orchestration, governance, and operability. Writing is how I pressure-test ideas and turn experience into reusable mental models — the kind that help senior engineers and architects ship reliably when systems become probabilistic and complex.
If you’re working on platform modernization, digital delivery, or operable AI/agent systems, feel free to browse my projects or my articles.
Most of my work lives where programs either stabilize and scale — or quietly drift into chaos:
- clarifying system boundaries and ownership
- turning integration ambiguity into explicit contracts
- shaping release processes that reduce surprises
- improving operability (rollback safety, observability, incident feedback loops)
Most background is grounded in computer science and engineering (BSc + MSc), and most of my professional work has lived inside complex, high-change environments where reliability and execution discipline matter more than theoretical correctness.
Background: BSc & MSc in Computer Science and Engineering.
How I think about architecture
I treat architecture as a delivery system, not a diagram. Good architecture:- makes change cheaper next month, not just “correct” today
- creates guardrails teams can follow under pressure
- designs for failure and recovery (not just the happy path)
- keeps truth boundaries clear: what must never be wrong, and how we prove it
Research & writing
Alongside client work, I’ve kept a long-running technical writing and research track.It started with classical ML fundamentals, moved through deep learning and RL failure modes, and has recently focused on agents as production systems: evaluation, tool safety, orchestration, governance, and operability. Writing is how I pressure-test ideas and turn experience into reusable mental models — the kind that help senior engineers and architects ship reliably when systems become probabilistic and complex.
Outside client work
I enjoy building small systems end-to-end: prototypes, simulators, tooling, and reference architectures — usually in service of reliability, automation, and learning.If you’re working on platform modernization, digital delivery, or operable AI/agent systems, feel free to browse my projects or my articles.