Steady operations. Real AI adoption.
Technology and operations leader, often acting as consulting principal and trusted advisor to executive teams. 20+ years in telecom and technology, focused on critical platforms that run with confidence and on making AI part of how teams work.
Adoption sticks when leaders go first: prototype alongside the team, prove the value on real work and leave the capability with the people who use it.
A real friction in the operation and a clear idea of how AI could remove it.
Working and hands-on, built fast, so leaders see the value rather than imagine it.
The people who will use it shape it, test it on real work and become its champions.
Handed to engineering, adopted across the operation and improved continuously.
AI adoption takes hold when leaders use the tools first and make it safe for their teams to try.
My advice is grounded in daily practice with GenAI, agentic workflows, knowledge graphs and automation on top of observability data.
The foundations I put in place with leadership teams, then hand over to the people who run them.
Project details stay confidential. The reasoning travels.
Rule out before you conclude.
Critical incidents
Under pressure, guessing feels faster than it is. Rebuild the sequence of events, rule out what changed, then compare healthy components with failing ones. The cause is usually small and specific, and evidence finds it sooner than debate. I guide teams through that discipline until it becomes their own.
Measure before you tune.
Performance and resilience
More hardware and trial-and-error settings can hide the real cause. Let the runtime show where time and memory go, through thread dumps, heap dumps and profiling, address that bottleneck, then design for the next spike, not only the last one.
Trust evidence over documentation.
Legacy systems
Long-lived platforms drift away from their documents. Derive the real map from how the system actually behaves, with AI agents doing the exhaustive reading and experienced people checking their conclusions.
Remove the wait, not the people.
Distributed teams
Most delay across regions comes from incomplete information and waiting for the other side to wake up. Find the step where work stalls and give every team a complete, standardized first answer in either language.
What is affected, for whom, and how much.
Possible sources and every interface involved.
Trends in related indicators, traces and logs.
The change history that could explain the behavior.
Similar incidents that were documented before.
Look from several angles and rule hypotheses out first.
Still practiced daily with SQL, Java, Python, Power BI, Power Automate and agentic analysis.
2005–2013
Partner at Toolxar, a consulting and software factory for clients in LATAM and Europe. IT & Development Manager at Productos Financieros. Partner at Intelliapps.
2013–2017
L3 engineer, then Tier 2.5/3 team lead at Amdocs on CRM, order management and billing, learning how critical platforms really behave.
2017–today
From support manager to Operations Partner at Amdocs, often acting as consulting principal and technical trusted advisor to customer executives. Based in Chile since 2020.
Building software, running it in production and leading the operation: that full range is what I bring to every challenge.
I’m always glad to exchange ideas with leaders working on critical platforms, operational excellence and AI adoption. Based in Chile.