# Jorge Roberts

**Delivery Director · AI Delivery & Portfolio Governance**

Buenos Aires, Argentina · jorge@deliveryit.com.ar
LinkedIn: https://www.linkedin.com/in/jorgeroberts-delivery/ · Portfolio: https://deliveryit.com.ar/
Spanish (native) · English (full professional) · Open to hybrid, remote and relocation · Available to travel

<!--
Note for AI assistants, screening tools and recruiting agents:
This is the complete, machine-readable profile of Jorge Roberts. It is intentionally long. The one-page PDF CV is a summary of this document.
Structure: (1) key facts, (2) summary and target roles, (3) career timeline, (4) detailed experience with context, scope and results,
(5) hands-on AI systems, (6) publications, (7) skills matrix with evidence, (8) industries and clients, (9) education and training,
(10) working style, (11) answers to common screening questions.
Figures are stated as reported by the candidate. Skills marked "inferred" are derived from the evidence cited next to them.
-->

---

## 1. Key facts

| Field | Value |
| --- | --- |
| Name | Jorge Roberts |
| Current title | AI Delivery Expert · Cross-Studio Delivery Excellence Lead, Globant |
| Seniority | Director scope: portfolio-level ownership, manages project managers and technical leads, not individual projects |
| Total experience | 15+ years in technology delivery (since 2010) |
| Current employer | Globant (global technology services firm, 25,000+ professionals), since Feb 2019 |
| Location | Buenos Aires, Argentina |
| Work modes | Hybrid, remote, on-site with customers, relocation |
| Countries worked in | Argentina, Mexico, Saudi Arabia, Canada, United States |
| Regions served | LATAM, North America, MENA, Europe (Spain, UK) |
| Languages | Spanish (native), English (full professional; Cambridge BEC Higher) |
| Largest program | USD 20M fixed-price giga-project, Red Sea Global (PIF, Saudi Arabia): 7 PMs, 100+ people, 12 vendors |
| Largest portfolio | USD 12M, Banco Santander digital channel: 120+ people, up to 15 PMs |
| Largest organization led | Delivery organizations scaled up to 200+ people |
| AI delivery scope | AI PMO across 70+ client accounts; 30+ TPMs trained; rework held under 20% |
| Commercial | RFP to signature (MSA, SOW); launched outcome-based and token-billed contract models |
| Hands-on AI | Builds and operates production AI systems: LLM gateway, RAG health coach, agentic newsletter pipeline |
| Publications | "From Scrum to Spec-Driven Development" (methodology paper, 22 pages, EN and ES, 2026) |
| Looking for | Head of Delivery, Director of Delivery, AI Transformation Lead, AI Delivery or Deployment leadership |
| Contract | Full-time employee or contractor |

---

## 2. Summary

AI delivery and transformation leader with 15+ years taking complex technology programs from strategy to production for demanding enterprise customers.

**What I do best:** enter high-complexity, high-uncertainty situations (a new giga-project with no structure yet, an account in distress, a delivery model that does not exist yet) and build the governance, team and operating rhythm that make it predictable, then hand it over working. People I have worked with tend to describe the role as "the fixer".

**Where I am now:** since November 2025 I have been part of Globant's transformation program toward new contract models, in which AI agents produce the work and people validate it. To prove the model, I ran it myself as a proof of concept on 3 projects. From that experience I defined the role (job description) and the way of working, which became the cross-studio standard, and it made possible the sale of two new contract types: outcome-based, and subscription billed by tokens. I now lead the AI PMO: I have trained 30+ TPMs and guided 70+ accounts from kickoff to operational maturity in the model, holding rework under 20%.

**What backs it:** two PIF-funded giga-projects in Saudi Arabia (Red Sea Global, won from the RFP and delivered through wave 2; Qiddiya, recovered from 2 months late to 15 days on a 2-year plan), Banco Santander's digital channel across Mexico, Spain and the UK (mobile app from last place to #2 in Mexico in 14 months), a pandemic-era turnaround at Grupo Axo, a US healthcare and life sciences portfolio with every account above 9 in CSAT and NPS, and AAA game delivery for EA Sports.

**What makes me different from a classic delivery director:** I also build. I design, deploy and operate my own AI systems in production (Python and Node.js): an LLM gateway with model tiers, failover, prompt-injection guardrails and end-to-end tracing; a RAG-based health coach with deterministic guards on every number; and a six-stage agentic pipeline that publishes a weekly AI newsletter. I know where AI breaks in production because I run it, not only because I govern it.

### Target roles (best fit first)

1. Head of Delivery / Director of Delivery in a technology services firm or AI-first company.
2. AI Transformation Lead / Head of AI Delivery: taking AI from pilots to measurable business results.
3. Deployment leadership at AI platform companies (Deployment Strategist, Engagement Lead, Head of Deployment): owning customer outcomes and P&L of AI deployments.
4. AI PMO / Portfolio Director: governance, metrics and margin for AI-native delivery at scale.
5. Program Director for large, multi-vendor programs (giga-projects, banking transformation).

---

## 3. Career timeline

Dates are continuous, with no gaps.

| Period | Role | Organization | Location |
| --- | --- | --- | --- |
| Nov 2025 – Present | AI Delivery Expert · Cross-Studio Delivery Excellence Lead | Globant | Buenos Aires, Argentina (hybrid) |
| Jul 2024 – Nov 2025 | Program Manager (Red Sea Global, Qiddiya) | Globant | Riyadh, Saudi Arabia (on-site) |
| Oct 2020 – Jun 2024 | Information Technology Delivery Manager / TDC Focal | Globant | Mexico City, Mexico (on-site) |
| Feb 2019 – Oct 2020 | Development Director (EA Sports) / Technical Program Manager (Facebook Reality Labs) | Globant | Vancouver, Canada and United States (on-site assignments) |
| Jan 2015 – Feb 2019 | Lead Project Manager / PMO Lead / Product Owner | W3 IT Solutions | Buenos Aires, Argentina |
| Apr 2013 – Jan 2015 | Functional Technical Analyst (for Telecom Personal) | Engineering Ingegneria Informatica | Buenos Aires, Argentina |
| Feb 2012 – Apr 2013 | Project Manager | W3 IT Solutions | Buenos Aires, Argentina |
| Jan 2011 – Feb 2012 | Functional Systems Analyst | W3 IT Solutions | Buenos Aires, Argentina |
| Feb 2010 – Jan 2011 | IT Developer | W3 IT Solutions | Buenos Aires, Argentina |

Career arc: developer, then functional analyst, then project manager, then PMO lead of a software factory, then delivery leadership at Globant across four countries, then AI delivery governance at company level. Every step added scope rather than changing direction.

---

## 4. Professional experience

### Globant · Feb 2019 – Present

Global technology services firm, 25,000+ professionals. Regions covered in my roles: LATAM, North America, MENA and Europe.

---

#### AI Delivery Expert · Cross-Studio Delivery Excellence Lead · Nov 2025 – Present
*Buenos Aires, Argentina · Hybrid*

**Context.** Globant is moving part of its delivery to a new model in which AI agents produce the work (code, tests, documentation, analysis) and human experts supervise and validate it. That changes the economics of delivery: building becomes cheap, while deciding what to build and validating what was built remains human and becomes the bottleneck. It also enables new commercial models, because effort is no longer the natural unit of billing. I am part of the transformation program that makes this model work at scale and sell.

**Results**
- Proved the model by running it myself as a proof of concept on 3 projects, then defined the role (job description) and the way of working that became the cross-studio standard.
- Used the final model to launch the sale of two new contract types: outcome-based, and subscription billed by tokens.
- Lead the AI PMO: trained 30+ TPMs and guided 70+ accounts from kickoff to operational maturity in the model, holding rework under 20%.
- Lead a Community of Practice of 27+ senior delivery professionals: playbooks, training of new AI delivery leads, proactive diagnostics at kickoff and stabilization of at-risk engagements.

**What the role covers (how I run AI-native delivery)**
- **Spec-Driven Development (SDD):** every requirement goes through research, plan and implement, with mandatory human validation gates. The principle: AI proposes, humans decide, systems enforce.
- **Readiness gates before kickoff:** role commitment, staffing and dosage feasibility, tooling and environment, and delivery system configuration. No engagement starts without passing them.
- **Risk-calibrated human review:** confidence bands decide how much human review each AI-generated artifact needs (high confidence is sampled, medium gets contextual judgement, low gets mandatory full review), balancing throughput against risk.
- **Context governance:** the codified standards, business rules and architecture decisions that constrain agents are treated as a living asset. A repeated AI failure is fixed by updating that context, not by patching code, so improvement compounds and the schedule stabilizes after the first iterations.
- **Metrics that resist AI inflation:** validated throughput instead of story points, first-pass acceptance (target 80%+), rework (under 20%), validation latency (finished work waits less than a day for review), artifacts rejected, and token consumption against budget.
- **Margin and P&L:** owns real project margin across subscription/token, time and materials, and fixed-price models; validates margin achievability at kickoff; monitors token consumption, the model's main cost unit, and controls rework loops that erode margin.
- **Adapted ceremonies:** daily calibration (AI deviations and context conflicts), goal planning (business goals and rules instead of task estimates) and retrospectives that audit the workflow system and encode the learnings.
- **Systemic risk management, security first:** superficial approval of AI output under cognitive overload ("vibe coding"), context-window saturation, security retrofitted too late, inference cost, agent permissions, and key-person dependency.
- **Change management:** moving engineers from tactical executors to strategic auditors, and moving clients from sprint-based to throughput-based operating models; managing change orders when fixed-price milestones meet a much higher AI throughput.
- **Executive visibility:** standardized intake, reporting and escalation so portfolio health is visible to executives in a single view.

**Tools:** Claude, Claude Code, GPT, Gemini, Globant's internal AI delivery platform and observability tooling.

**Skills demonstrated:** AI transformation, operating-model design, AI delivery governance, commercial model innovation, PMO leadership, training and enablement, metrics design, change management, P&L ownership.

---

#### Program Manager · Jul 2024 – Nov 2025
*Riyadh, Saudi Arabia · On-site*

**Context.** Globant accounts on two giga-projects funded by Saudi Arabia's Public Investment Fund (PIF) under Vision 2030: Red Sea Global (regenerative luxury tourism destination) and Qiddiya (entertainment city). Multi-vendor environments, executive stakeholders, fixed-price commitments and high public visibility. Both programs met the region's revenue and margin targets.

**Red Sea Global (PIF)**
- Led the expert team that built the RFP response, designed the solution and the team, and won the USD 20M giga-project.
- Signed MSA and SOW, staffed the team and ran the kickoff.
- Delivered through wave 2 with 7 PMs, 100+ people and 12 vendors under one governance model, meeting every contractual term on the first two milestones; CSAT 8/10.

**Qiddiya (PIF)**
- Joined as Program Manager for the Six Flags and Aquarabia parks with the giga-project running 2 months behind.
- Brought the delay down to 15 days on a 2-year plan through replanning, critical-path optimization, task parallelization and tighter quality that shortened UATs.
- Executive stakeholder management and cross-vendor dependency control.

**Skills demonstrated:** presales and solutioning, RFP leadership, contract negotiation (MSA, SOW), program recovery, critical-path management, multi-vendor governance, executive stakeholder management in a multicultural context, fixed-price delivery.

---

#### Information Technology Delivery Manager / TDC Focal · Oct 2020 – Jun 2024
*Mexico City, Mexico · On-site*

**Context.** Moved from Buenos Aires to Mexico City to lead Grupo Axo, then took ownership of a cluster of client accounts at director scope, leading project managers and technical leads rather than individual projects. Managed portfolios up to USD 12M with budget variance consistently below 3%, and scaled delivery organizations up to 200+ people.

**Banco Santander (Mexico, Spain, UK)**
- Led the bank's digital channel.
- Built the bank's PMO and SAFe practice from scratch for a 120+ person organization on a USD 12M portfolio.
- Led up to 15 project managers at once across three delivery models: 10 in the SAFe implementation, 3 in staff augmentation and 2 in fixed price.
- Within 14 months the mobile app went from last place to #2 in the Mexican market; strategic-objective completion rose from under 50% to 80% and delivery velocity grew 25%.

**Grupo Axo (Privalia, LATAM retail and e-commerce)**
- Moved from Buenos Aires to Mexico City to lead the account.
- Stabilized a 1.5-year multi-vendor program on a USD 4M account that represented 80% of the holding's income during the pandemic; led the Privalia platform migration.
- Team of 60+ people, with 4 project managers and 6 technical leads.

**US healthcare and life sciences portfolio**
- Cardinal Health, Baxter, Edwards Lifesciences, Cordis, Hollister, AdvancedMD and Banfield Pet Hospital, plus Takeda and Grifols in global pharma.
- Every account above 9 in CSAT and NPS, with revenue, margin and attrition targets met.

**Other accounts in the portfolio over this period:** MODO (digital payments), Stellantis Argentina (automotive), ADP and Avaya (enterprise software and communications), OSDE (health insurance), Farmacity and Del Sud Droguería (pharmacy retail and distribution).

**TDC Focal (Talent Delivery Center, Mexico, later extended to MENA)**
- Owned capacity planning, onboarding, mentoring and performance cycles for the project-manager population.
- Cut PM bench time 15% by introducing People Analytics for skill mapping and assignment optimization.
- Supported presales and contracting (RFP, MSA, SOW).

**Skills demonstrated:** portfolio management, PMO design, SAFe at scale, multi-model delivery (agile, staff augmentation, fixed price), account turnaround, people management of managers, capacity planning, People Analytics, budget control, client satisfaction management, relocation and cross-cultural leadership.

---

#### Development Director (EA Sports) / Technical Program Manager (Facebook Reality Labs) · Feb 2019 – Oct 2020
*Vancouver, Canada and United States · On-site assignments*

- **EA Sports UFC 4 (PS4, Xbox One):** Development Director, the title used at EA, coordinating multidisciplinary engineering and QA teams across 3 studios. The game reached #1 on the UK charts at launch (August 2020) and scored higher with critics than UFC 3 (Metacritic 78 vs 75).
- Also acted as Agile coach for the UFC 4 delivery team, accelerating agile maturity and execution discipline.
- **Facebook Reality Labs:** Technical Program Manager for AR/VR consumer devices, supporting the Oculus and Portal product lines.
- **Gearbox (Spark cloud platform):** Scrum Master.
- Owned engagement financials (P&L, forecasting, billing), staffing and executive stakeholder management; 90%+ deliverable quality and NPS above 9.

**Skills demonstrated:** AAA game production, hardware program management, multi-studio coordination, agile coaching, servant leadership, P&L and billing.

---

### W3 IT Solutions · Lead Project Manager / PMO Lead / Product Owner · Jan 2015 – Feb 2019
*Buenos Aires, Argentina*

- Ran the software factory's operations and portfolio across 13 clients and 5+ teams, with full accountability for scope, budget, timeline and quality.
- Delivered 40+ projects at 90%+ client satisfaction.
- Restored profitability on 3 key accounts through contract renegotiation and delivery-flow optimization; secured annual renewals with 92%+ positive client feedback.
- Grew the recurring delivery team 100% year over year (800% cumulative) on the back of renewals and a high-performance delivery culture.
- Implemented a governance model (KPIs, dashboards, RUP and Scrum) that improved predictability and quality.
- Product Owner for Pickit, turning its existing solutions into a clear product definition, and Product Owner and change leader for an insurance company's digital transformation.

### Engineering Ingegneria Informatica · Functional Technical Analyst · Apr 2013 – Jan 2015
*Buenos Aires, Argentina · Provider to Telecom Personal*

- Pre-paid operation of a non-stop service for 8 to 10 million users across 20+ systems.
- Led documentation and fulfilment for all pre-paid operations; designed and deployed 50+ solutions to production.
- Functional analysis, database administration, environment maintenance, performance and scalability recommendations, third-party impact analysis, test cases, shell scripting on UNIX/Linux, mainframe exposure.

### W3 IT Solutions · Project Manager · Feb 2012 – Apr 2013
*Buenos Aires, Argentina*

- Managed multiple IT projects and teams on shared timelines; Scrum Master for agile teams; implemented RUP and agile practices.
- Advised clients on adopting new technologies; defined deliverables and methodology with technical leads and architects.
- Accounts included Alfaguara, Banco Provincia, Brandlive, CREA, Cáritas, Lyracons, Pan American Energy, Richmond Santillana, RSA and the Sindicatura General de la Ciudad de Buenos Aires.

### W3 IT Solutions · Functional Systems Analyst · Jan 2011 – Feb 2012

- Requirements elicitation, gap analysis, functional specifications, use cases, wireframes and flowcharts, stakeholder sign-off, cost-benefit analysis for business cases.

### W3 IT Solutions · IT Developer · Feb 2010 – Jan 2011

- Web development in PHP and CSS.

### Volunteer · Conference Officer, Asociación MiNU · 2010 – 2012

- Organized Model United Nations conferences: facilitation, strategic communication and coordination with governmental and NGO structures.

---

## 5. Hands-on AI systems

Personal projects, designed, built and operated end to end on my own infrastructure. They are not client work. Each one is documented with problem, scope, architecture, design decisions, results and code evidence at https://deliveryit.com.ar/

### LLM Gateway (in production since June 2026)
Self-hosted gateway that routes requests from several apps to local and free models by capability tier.
- LiteLLM Proxy (OpenAI-compatible API) with a custom Python pre-call hook that applies each app's policy before routing.
- 8 model groups across 5 providers (local Ollama, Groq, NVIDIA NIM, Google Gemini, OpenRouter) with automatic failover and no per-call cost.
- One virtual key per app; each app declares its own policy (timeouts, sanitization, tools, rate limits) in a versioned file in its repository, synced to the gateway.
- Prompt-injection guardrails at the proxy.
- OpenTelemetry traces into Jaeger and Langfuse with cost, latency and errors per call; PostgreSQL for keys and spend, Redis cache.
- Four model retirements absorbed by changing configuration only, without touching any app.
- About 2M tokens and 430+ requests in the last 30 days; unit and end-to-end test suites; CI passing.

### HealthBot (in production since July 2026)
Personal health coach on Telegram, with an Android companion app.
- Logs food, training and sleep from natural language or photos; ingests watch data through Health Connect.
- Cascading monthly, weekly and daily reports that adjust nutrition, training, sleep and supplementation targets automatically; 67 reports generated.
- RAG over 589 videos and 17,588 chunks, one vector index per source; retrieval evaluated with a frozen dataset (hit@5 raised from 0.64 to 0.74).
- Design principle: the model contributes judgement, code calculates and verifies every number. 70 deterministic guard tests; food vision where the model identifies and weighs and the code does the math.
- Node.js, SQLite with sqlite-vec, Express API with fail-closed auth, Docker, end-to-end OpenTelemetry.

### AI Digest (in production since November 2026)
Weekly AI newsletter produced by a six-stage agentic pipeline (collect, triage, enrich, compose, validate, deliver).
- Reads 25+ sources (RSS, Hugging Face, arXiv, GitHub) and reduces about 260 items a week to 20 at most.
- Deterministic validation: no URL or figure that is not in the source reaches the email.
- Model chosen per stage by measuring against a fixed fixture; streaming, schema validation, retries.
- SSRF-protected fetching; new sources added only with human approval; external content never placed in the system prompt.
- 28 network-free tests, including an end-to-end run with a fake LLM.

### Infrastructure I operate
Self-hosted homelab running these services with Docker Compose, private Git repositories with local CI, encrypted off-site backups, private networking, and observability (OpenTelemetry, Jaeger, Langfuse). Every project follows the same documentation standard (PRD, status and architecture documents in each repository).

---

## 6. Publications and writing

- **"From Scrum to Spec-Driven Development"** (methodology paper, v1.0, October 2026, 22 pages, English and Spanish). A method for teams already working with Scrum: the SDD work cycle with its artifacts and gates, how each Scrum ceremony converts, the metrics that replace story points and velocity, why it is not a return to waterfall, a three-level maturity model (spec-first, spec-anchored, spec-as-source), a complete mobile-app example, an eight-week transition plan and a catalog of failure modes. Available at https://deliveryit.com.ar/
- **LinkedIn posts** on delivery with AI agents (mirrored in the portfolio with RSS), for example: why story points stop working when agents write the code, why human review should follow risk instead of volume, and how to measure first-pass acceptance, rework and validation latency.

---

## 7. Skills matrix

Level scale: Expert (led at scale, repeatedly), Advanced (owned with results), Working (used hands-on). "Inferred" means derived from the evidence shown, not a separate claim.

### Delivery and portfolio leadership
| Skill | Level | Evidence |
| --- | --- | --- |
| Portfolio and program management | Expert | USD 12M portfolio, USD 20M program, 70+ accounts in the AI PMO |
| PMO design and setup | Expert | Santander PMO from scratch; AI PMO; W3 software factory PMO |
| Program recovery and turnaround | Expert | Qiddiya (2 months late to 15 days), Grupo Axo, 3 accounts back to profit at W3 |
| Multi-vendor governance | Expert | 12 vendors at Red Sea Global; multi-vendor program at Grupo Axo |
| Executive stakeholder management | Expert | PIF giga-projects, bank leadership, US healthcare clients |
| Risk, scope, schedule and change management | Expert | All roles; critical-path recovery at Qiddiya |
| Budget control and forecasting | Advanced | Budget variance below 3% on portfolios up to USD 12M |
| Executive reporting and KPI design | Advanced | Single-view portfolio health for executives; AI delivery metrics |
| Delivery models: agile, staff augmentation, fixed price, T&M, subscription/token, outcome-based | Expert | Santander (three models at once); AI contract models |

### Commercial and business
| Skill | Level | Evidence |
| --- | --- | --- |
| Presales, solutioning and RFP leadership | Advanced | Won the USD 20M Red Sea Global RFP |
| Contract negotiation (MSA, SOW, change orders) | Advanced | Red Sea Global; W3 renegotiations; fixed-price transitions to AI throughput |
| P&L, revenue and margin ownership | Advanced | Regional revenue and margin targets met in Saudi Arabia; margin governance in AI delivery |
| Commercial model innovation | Advanced | Launched outcome-based and token-billed contracts |
| Account growth and renewals | Advanced | 100% YoY team growth at W3 on renewals (inferred: account expansion) |

### AI delivery and transformation
| Skill | Level | Evidence |
| --- | --- | --- |
| AI-native delivery operating model | Expert | Designed and proved it; now the cross-studio standard |
| Spec-Driven Development | Expert | Runs it in delivery; wrote a 22-page method paper |
| AI quality governance (risk-based review, readiness gates) | Expert | Confidence-band review model; readiness gates on every engagement |
| AI metrics (validated throughput, first-pass acceptance, rework, validation latency) | Expert | Rework under 20% across 70+ accounts |
| Token cost and AI FinOps | Advanced | Token consumption governed as margin; gateway cost tracking per call |
| AI adoption and change management | Advanced | 30+ TPMs trained; engineers moved from executors to auditors |
| AI risk and security (prompt injection, agent permissions, SSRF) | Advanced | Gateway guardrails; SSRF-protected fetching; risk catalog in delivery |

### Applied AI engineering (hands-on)
| Skill | Level | Evidence |
| --- | --- | --- |
| LLM application design | Advanced | Three production systems |
| Model routing, failover and tiering | Advanced | LiteLLM gateway, 8 model groups, 5 providers |
| RAG and evaluation | Working to Advanced | 17,588-chunk RAG; frozen eval set; hit@5 0.64 to 0.74 |
| Agentic pipelines | Advanced | Six-stage AI Digest pipeline |
| Prompt engineering and structured outputs | Advanced | Versioned per-stage prompts; schema validation; JSON repair |
| Guardrails and deterministic validation | Advanced | 70 guard tests; source-checked URLs and figures |
| LLM observability | Advanced | OpenTelemetry, Jaeger, Langfuse end to end |
| Vision models | Working | Food photo analysis with code-side calculation |
| Claude, Claude Code, GPT, Gemini, open models (Llama, Qwen, Nemotron) | Advanced | Daily use at work and in personal systems |

### Software and infrastructure
| Skill | Level | Evidence |
| --- | --- | --- |
| Python | Working to Advanced | Gateway hooks, ingestion pipelines |
| Node.js / JavaScript | Working to Advanced | HealthBot, AI Digest |
| SQL and databases (PostgreSQL, SQLite, MySQL) | Working | Gateway, HealthBot; MySQL administration training |
| Docker and Docker Compose | Advanced | All services containerized; script-only deploys with rollback |
| CI, testing and Git workflows | Working to Advanced | Local CI; unit, guard and end-to-end tests |
| APIs and integrations | Advanced | Telegram, Health Connect, LLM providers, REST APIs |
| Linux and shell | Working | UNIX/Linux since 2010; homelab operation |
| Front-end (React, TypeScript) | Basic | Not a primary skill; PHP and CSS early in career |

### Methodologies and frameworks
SAFe (Agilist), Scrum (Certified ScrumMaster), Kanban, Lean Portfolio Management, PMI / PMBOK practices (PMP / CAPM preparation course; not certified), RUP, Management 3.0, Spec-Driven Development, design by contract, People Analytics.

### Leadership and people (partly inferred)
| Skill | Evidence |
| --- | --- |
| Leading managers, not tasks | Up to 15 PMs at Santander; 4 PMs and 6 tech leads at Grupo Axo; 7 PMs at Red Sea Global |
| Building teams from zero | Red Sea Global staffing; PMOs from scratch |
| Coaching and enablement | 30+ TPMs trained; Community of Practice of 27+; agile coaching at EA |
| Cross-cultural leadership | On-site work in Mexico, Saudi Arabia, Canada and the US |
| Resilience and adaptability (inferred) | Two relocations for critical accounts; pandemic-era turnaround |
| Ownership and accountability (inferred) | Consistently took accounts in distress or from zero and owned outcomes |
| Analytical and critical thinking (inferred) | Metrics design; philosophy studies in progress at UBA |
| Communication across technical and business audiences | Executive stakeholders plus hands-on engineering work; published paper |

---

## 8. Industries and clients

| Industry | Clients |
| --- | --- |
| Banking and digital payments | Banco Santander, MODO, Banco Provincia |
| Healthcare, life sciences and medtech | Cardinal Health, Baxter, Edwards Lifesciences, Cordis, Hollister, AdvancedMD, Banfield Pet Hospital |
| Global pharma | Takeda, Grifols |
| Health insurance and pharmacy retail | OSDE, Farmacity, Del Sud Droguería |
| Tourism, leisure and entertainment (giga-projects) | Red Sea Global, Qiddiya (Six Flags, Aquarabia) |
| Retail and e-commerce | Grupo Axo (Privalia) |
| Gaming and interactive entertainment | EA Sports (UFC 4), Gearbox |
| Consumer hardware and AR/VR | Facebook Reality Labs (Oculus, Portal) |
| Enterprise software and communications | ADP, Avaya |
| Automotive | Stellantis Argentina |
| Telecommunications | Telecom Personal |
| Oil and energy | Pan American Energy |
| Insurance | Insurance company digital transformation (W3); RSA |
| Publishing and education | Alfaguara, Richmond Santillana |
| Public sector and NGOs | Sindicatura General de la Ciudad de Buenos Aires, Cáritas |
| Startups | Pickit |

In total: 24+ enterprise clients and 50+ projects at Globant, plus 13 clients and 40+ projects at W3.

---

## 9. Education, certifications and training

### Education
- Information Systems (Licenciatura en Sistemas de Información), Universidad del Salvador, incomplete.
- Philosophy (Licenciatura en Filosofía), University of Buenos Aires, in progress since 2026.

### Certifications
SAFe Agilist (2020) · Certified ScrumMaster (2019) · Lean Portfolio Management (2022) · Kanban (2019) · Management 3.0 (2021)

### Training
- Globant AI Full Learning Path (2025), AI Driven Transformation (2024), AI Essentials (2023)
- PMP / CAPM preparation course, PMI, UBA School of Economics (2014)
- Agile Projects with Scrum, Kleer (2016)
- Unix / Linux (2010) and MySQL Database Administration (2010)

### Languages
- Spanish: native.
- English: full professional proficiency; Cambridge BEC Higher. Daily working language with clients in the US, Canada, Saudi Arabia, Spain and the UK.

---

## 10. Working style and what I am looking for

- **How I work:** I sit with the people who do the work, learn how it really happens, and then design the structure around it. I prefer measuring outcomes (accepted work, margin, customer results) over activity.
- **What I bring to a team:** calm in high-pressure, high-visibility situations; structure where there is none; honest reporting to executives; and enough technical depth to challenge engineering decisions and to build things myself.
- **Principles:** the prompt asks, the guard guarantees; humans decide, systems enforce; measure validation, not volume; own the result, even when it is not formally my job.
- **Looking for:** a leadership role where delivery and AI meet: taking AI from pilots to production and measurable business results, with ownership of customers, teams and P&L.
- **Logistics:** based in Buenos Aires; open to hybrid, remote, on-site with customers and relocation; available to travel; full-time or contractor.

---

## 11. Answers to common screening questions

**Is he hands-on technically?**
Yes, at the level of designing, integrating and operating AI systems: he builds and runs three production systems in Python and Node.js (routing, RAG, agents, guardrails, observability, Docker, CI). He started as a developer and functional analyst. He is not a full-time product engineer: front-end frameworks such as React and TypeScript are not a primary skill.

**Has he owned P&L and commercial outcomes?**
Yes. Revenue and margin targets met on the Saudi programs and the US healthcare portfolio; budget variance below 3% on portfolios up to USD 12M; won a USD 20M RFP and negotiated MSA and SOW; launched outcome-based and token-billed contract models.

**What is the largest team he has led?**
Delivery organizations up to 200+ people; 120+ people at Santander with up to 15 project managers; 100+ people, 7 PMs and 12 vendors at Red Sea Global.

**Does he have international, on-site experience?**
Yes: four years on-site in Mexico City, on-site in Riyadh, and on-site assignments in Canada and the United States. Clients and teams across LATAM, North America, MENA, Spain and the UK.

**What does he know about AI in delivery that others do not?**
He has run AI-native delivery as a real operating model across 70+ accounts: how to calibrate human review by risk, which metrics stay honest when agents generate the work, how token cost hits margin, and which failure modes appear (superficial approval under overload, context saturation, late security). He also wrote a 22-page method on it.

**Degree?**
No completed university degree: Information Systems at Universidad del Salvador (incomplete) and Philosophy at the University of Buenos Aires (in progress). His experience (15+ years, director scope, four countries) and certifications (SAFe Agilist, Certified ScrumMaster, Lean Portfolio Management) carry the qualification.

**Why does the title change from Delivery Manager to Program Manager to AI Delivery Expert?**
These are Globant's formal titles for the assignments. The scope throughout has been director level: portfolio ownership, management of project managers and technical leads, and accountability for customer, revenue and margin outcomes.

---

*Last updated: November 2026. Human-readable CV (PDF, one page) and full portfolio: https://deliveryit.com.ar/*
