Software engineer · Solution architecture · Buenos Aires
I am Gabriel Paz, a hands-on software engineer and solution architect withfive years of verifiable professional experience. I work where AI workflows, distributed systems, product delivery, and technical teams need a shared operating model — not another layer of hidden complexity.
Selected directions
I am at my best when a team has to make a technical system clearer, safer, more useful, and easier to evolve at the same time.
Architecture and workflow design for sensitive-domain assistants, with clear action boundaries, traceability, observability, and human review paths.
API boundaries, permissions, workflow states, auditability, and practical delivery decisions for business-critical systems.
I build the context, evidence, and orchestration layers that help teams use AI without losing technical judgment or accountability.
Public projects and tools
Each one solves a different part of the same problem: make context clearer, decisions safer, and technical work easier to carry forward. They are public, evidence-bounded views — not confidential case studies in disguise.
Product ecosystem
Contedi is where I explore products that turn everyday capture into retraceable context. Tedi Notas gives that work a public writing surface; Tedi brings notes, sources, recall, uncertainty, and user correction together so personal knowledge is not treated as a black box.
A public product narrative. It intentionally excludes private users, customer data, and unverified product claims.
Visit ContediTeam operating model
The wiki holds shared reality: decisions, constraints, evidence, and what is still open. The execution layer turns that context into bounded work with ownership, stop conditions, review points, and human judgment where it matters.
A public method and criteria case. No client, team, or operating details are disclosed.
Explore mi-wikiDeveloper tooling
A local CLI for finding relevant code, reading exact ranges, following repository relationships, and starting from canonical documentation. It replaces blind repository discovery with focused, inspectable context.
Public repository material; private project context and infrastructure details stay out of this profile.
Explore mi-lspData tooling
A terminal-first tool for PostgreSQL and SQL Server that keeps connection configuration explicit, supports structured output, and makes writes and migrations deliberate rather than accidental.
Public repository material. It describes the tool's safeguards, not access to any production database.
Explore db-cliBrowse the wider body of public work on GitHub.
Teaching the new way of working
I help teams move beyond individual prompting toward coordinated AI-assisted delivery: shared context, explicit ownership, evidence, and decisions that remain inspectable. The goal is not more automation; it is a team that can make better technical choices together.
It gives decisions, constraints, evidence, and open questions a durable home. Instead of rebuilding the same context in every conversation, people can inspect it, challenge it, and extend it together.
It gives each task an explicit scope, owner, stop condition, evidence route, and human decision point. AI-assisted delivery becomes faster without turning judgment, quality, or accountability invisible.
This is what I teach: move from isolated AI use toward a shared practice where people can find context, challenge decisions, delegate safely, and learn from the evidence the work produces.
How I work
I frame the system around the real constraints first: users, reliability, privacy, integration boundaries, and the decisions that must remain visible.
I stay close to the implementation layer: C#/.NET, APIs, distributed workflows, observability, and AI-assisted development where it is genuinely useful.
I help turn a good technical direction into a working rhythm: shared context, smaller decisions, explicit ownership, and room for people to learn.
Let’s talk