Technical project management · Delivery systems · AI-assisted implementation
I turn fragmented workflows, technical complexity, and operational ambiguity into governed systems teams can actually run—from CRM and product delivery to automation, reporting, UAT, and release readiness.
Technical Project Manager · Delivery Systems Architect · AI-Enabled Operations & Integration Lead
15+ years across software delivery, operations transformation, consulting, and global remote teams · Barranquilla, Colombia
The operating model moves from business goal to source of truth, workflow, automation, evidence, and finally a clear decision.
My role is to understand what should be built, why it is correct, how the systems should interact, what can go wrong, and what evidence proves the result is ready.
I work between business, product, engineering, QA, operations, and leadership—turning decisions into buildable work and technical complexity into clear business tradeoffs.
I use AI-assisted coding as an implementation engine, not a substitute for judgment. I own the constraints, architecture decisions, acceptance criteria, risk controls, validation, and final direction.
The operating perspective
“The system comes before the tool. If ownership, truth, and evidence are unclear, automation only makes the confusion move faster.”Read my approach ↗
Metrics belong to specific engagements and are shown with context in the related case studies.
Caul Group / Sympli Mortgage
A governed operating layer across systems that already held different parts of the truth—designed to protect borrower communications, reduce duplicate data, and give leadership useful visibility.
Read the case study ↗01Canonical truth stayed where the business already governed it.
02Borrower-facing uncertainty failed safely and escalated internally.
03Automation execution cost reduced by approximately 55% in the engagement.
Each system begins with the operating truth: who owns what, how work moves, what can fail, and what evidence earns the next decision.
TH3RD GROUP
One operating model across commercial and delivery systems.
Commercial truth separated from operational truth.
Open case file ↗FastVideo
Product architecture and launch operations moving as one.
Six-stage AI video workflow with governed handoffs.
Open case file ↗Jane’s Weather / ACM
Make release readiness visible across product, QA, and APIs.
75% shorter PR review cycle in the engagement.
Open case file ↗Jira
Make workflow status reflect engineering reality.
Evidence-linked movement from intake to release.
Open case file ↗Victus Global
Connect intake, approvals, delivery, and visibility.
Lifecycle automation built around business rules.
Open case file ↗Tools can accelerate implementation. They cannot decide which truth matters, where the risk sits, or what proof makes the result safe to use.
Turn business ambiguity into requirements, acceptance criteria, and explicit tradeoffs.
Weather delivery: API and UAT expectations became testable work.Define system boundaries, sources of truth, workflow logic, and integration rules.
Sympli: ARIVE remained authoritative while other systems enriched safely.Build controls for risk, data quality, evidence, permissions, and release readiness.
Jira: status movement required proof from the delivery flow.Coordinate implementation, validate outcomes, recover responsibly, and reduce operating drag.
Mortgage operations: automation was consolidated around value and plan limits.I work closely with engineering across APIs, repositories, pull requests, dependencies, environments, QA, deployment, and release flow. I own the problem definition, architecture constraints, workflow logic, requirements, tradeoffs, testing expectations, and final acceptance.
Engineers own engineering. AI tools can accelerate implementation. My responsibility is to make sure the right system is being built and that evidence—not momentum—earns release.
The through-line is not a job title. It is making complex work legible, governable, and ready to move.
Controls, service, ownership, and measurable cycle improvement.
CRM, ERP-lite systems, automation, and role-based visibility.
Requirements, APIs, QA, UAT, release readiness, and global teams.
Architecture-led implementation with validation and guardrails.
A useful place to begin
When work is moving but control is missing, I map the truth, define the operating model, and build the structure that helps the team execute without guessing.