Skip to content

Technical delivery leader at Globant — enterprise AI deployment.

SantiagoLópezZavaletta

I run enterprise AI and software delivery—from staffing and P&L to risk, adoption, and the client decisions that follow.

My career started in infrastructure. Today I work between clients, executives, and technical teams across the United States and Latin America.

Enterprise AI deployment is the work: choosing the use case, running the pilot, supporting the people whose work changes, and making the scale-or-stop call on evidence—across three enterprise implementations and a $5M AI account.

Read the workDownload résumé

Evidence

Current / Next

What I do now. Where I’m taking it next.

Current practice

I manage the scope, budget, people, risk, and the work itself.

At Globant, I own scope, budget, staffing, timelines, dependencies, risk, client communication, and the Scrum cadence for distributed software teams. My portfolio work has ranged from $700k to $5M, with cross-functional teams of up to 36 people. I use financial and delivery metrics to make tradeoffs while the team still has room to act.

  1. Scope, plans, budgets, and staffing
  2. Scrum cadence, dependencies, and risk
  3. Client communication, quality, and delivery signals

Where this goes

The AI deployment work I lead, and where I’m taking it.

AI can produce a persuasive demo before a company knows how—or whether—to adopt it. The harder work is choosing the right use case, handling data and security constraints, supporting the people whose work will change, and deciding whether the evidence is strong enough to continue. That is the work I do, and the work I want more of.

In progress

I’m pursuing the Claude Certified Architect certification and building hands-on depth through delivery tooling, automation, self-hosted infrastructure, and experiments with agent workflows.

01 / Selected work

Selected work.

Three examples: what was difficult, what I owned, and what changed.

01

AI delivery

Globant · Oct 2025–present

Improving margin and reducing overhead on a $5M AI account.

A multidisciplinary AI account had margin pressure and too much recurring delivery work happening by hand. I own staffing, capacity, and P&L for a 20-person delivery team within a 40-person account, run Scrum and governance, and automate recurring work with Claude and Jira. Account margin moved from 40% to 45%.

Result signal

Five percentage points of margin improvement.

Context
AI delivery across data science, engineering, front-end, and DevSecOps.
Constraint
Margin pressure and recurring operational overhead.
Work
Staffing mix, capacity planning, RAID governance, Scrum, and workflow automation.

02

Digital Twin Studio

Globant · Dec 2024–Oct 2025

Building a new Digital Twin capability through six POCs.

Globant wanted a Digital Twin capability in a domain that was new to the team. I led a five-person LATAM POD, delivered a six-month, $90k fixed-price factory twin, worked with external partners, and supported presales. The first POC took roughly five weeks; the studio went on to deliver six.

Context
A new real-time 3D capability with no existing delivery model.
Constraint
An unfamiliar domain, external partners, and a fixed-price commitment.
Work
Team design, vendor coordination, POC delivery, and a repeatable studio model.

Result signalSix POCs, roughly five weeks to the first, and a $90k fixed-price delivery.

03

M&A and corporate development

Globant · Jun 2022–Mar 2023

Coordinating four M&A programs across four countries.

I coordinated due diligence across three Latin American acquisitions and the post-merger integration of a European firm. The work covered legal, marketing, IT, and change-management risk across approximately 940 people.

Context
Simultaneous due diligence and post-merger work.
Constraint
Different countries, functions, and decision owners.
Work
Cross-functional coordination, risk tracking, stakeholder communication, and integration planning.

Result signalFour deals covering approximately 940 people.

Running an AI deployment

Before a pilot starts, I want five things written down.

  1. 01The business problemWhat is costly, slow, risky, or otherwise worth changing?
  2. 02The usersWhose workflow needs to change, and what support will they need?
  3. 03The constraintsWhat do data access, security, integrations, time, and competing tools allow?
  4. 04The success signalWhat evidence would be credible enough to act on?
  5. 05The decisionWhat will the evidence allow the customer to stop, fix, expand, or buy?

Working brief / v0.1

This is the brief I work from: what the delivery team and the client both have to agree on before the first sprint, drawn from the implementations I have run.

Why this page

This gives the client and delivery team a shared answer to two questions: what are we trying to prove, and what will we do with the result?

02 / Built around the work

Delivery systems I build.

These tools keep scope and evidence visible, so the next delivery decision doesn't depend on memory.

Delivery system

SOW Intake

SOW Intake turns a contract into a cited delivery baseline that people and agents can use. Missing evidence is marked as missing instead of being filled with a plausible answer.

View the source on GitHub (opens in new tab)

Behind the experiments

I also run the infrastructure behind my own experiments: Docker, Caddy, Tailscale, n8n, project tracking, and personal agents on a self-hosted VPS.

Scope Sentinel / Walkthrough

Follow a client request from SOW clause to next step.

Choose a fictional client request. The walkthrough checks it against a sample SOW, cites the clause, sizes the impact, and drafts the next step.

Sample baseline

Customer Portal / Phase 1

Fixed price
Deliverables
D1 — SSO and role-based access, up to three rolesD2 — Read-only account dashboard, four widgetsD3 — Create and view your own support tickets
Exclusions
Native mobile applicationsThird-party integrations beyond SSOLegacy data migration
Assumption
One round of UAT before go-live

Pick an incoming client request

IN SCOPEClassification

Citation / Deliverable D3

“Create and view your own support tickets.”

Reasoning

Viewing an existing ticket is already part of the agreed deliverable.

Size

No additional size

Next step

Keep the request in the current delivery baseline.

Fictional sample data. The real skill works inside Claude against an actual SOW and request; citations and draft actions remain subject to human review.

03 / Working set

Tools, grouped by how I use them.

Tools matter when they shorten a feedback loop, make a decision easier to trace, or remove work a team should not be doing by hand.

Run the work

I use these for scope, backlogs, delivery decisions, and shared context.

  • Jira
  • Power BI
  • Linear
  • Figma

Build and automate

The tools I use to draft, test, and remove repeatable delivery work.

  • Claude
  • ChatGPT / Codex
  • Gemini
  • n8n

Ship and operate

A practical stack for scripts, source control, containers, and the infrastructure behind my experiments.

  • Python
  • Bash / PowerShell
  • GitHub
  • Docker

Explore AI systems

Active learning: I use these to understand orchestration, observability, evaluation, and retrieval more deeply.

  • Mastra
  • LangGraph
  • Langfuse
  • pgvector

04 / Experience

How I got here.

Two engagements ran concurrently with my role at Globant, and are marked below.

Download résumé

  1. 01

    2021–present

    Globant

    Technical Project Manager

    End-to-end delivery, forecasting, and P&L for US and LATAM digital transformation portfolios from $700k to $5M, with cross-functional teams of up to 36. Launched Globant’s Digital Twin Studio and standardized AI-assisted delivery reporting across the account.

  2. 02

    2026–present

    Concurrent with Globant

    ZN Love

    Fractional Project Manager

    Client relationships and delivery health across a three-project portfolio spanning entertainment, workforce technology, and creative production. Building the PMO from the ground up: presales discovery, SOW, governance, and change control.

  3. 03

    2025–2026

    Concurrent with Globant

    Blue Crab Consulting

    Engagement Manager

    Led functional and technical consultants across three enterprise Eightfold AI implementations — a global beverage leader, a leading LATAM retailer, and a top US research university — delivering on-time go-lives and full adoption, and scaling one deployment from its Mexico launch across additional LATAM markets.

  4. 04

    2018–2021

    ExxonMobil

    System Administrator

    Enterprise infrastructure, security, reporting, and service transition.

  5. 05

    2014–2018

    gA

    Technical Project Lead

    Enterprise migration and performance-testing programs in life sciences.

05 / Contact

I’m looking for my next role in AI deployment.

If you need a technical delivery leader who can own the delivery around an AI pilot—scope, people, risk, budget, and client decisions—I’d like to talk.

Email SantiagoConnect on LinkedIn (opens in new tab)View GitHub (opens in new tab)