Valencia, Spain · 7+ years in supply chain and operations

Facundo Bertua

Supply chain leader who ships AI into operations.

Most recently at Amazon and DHL. I lead a team of 4 engineers that creates value in the Airbus supply chain, and I build the AI tools behind that work: cost forecasting with data-quality checks for the Airbus account, and savings sizing for new-client RFQs.

Actual · DHL€3M+annual savings delivered by the team I lead in the Airbus supply chain
Actual · 2024 and 2025−5%transport cost per year, from the Cost Performance Management service I developed
Actual · in production18+Airbus transport managers served by the AI forecasting tool I built

Email LinkedIn

About me

I'm an industrial engineer who ended up where operations meet software. At Amazon I learned to plan operations at scale; at DHL I lead the Value Creation team for the Airbus supply chain and build the AI tools behind that work.

What I'm looking for: a strategic role in supply chain, operations or business development, or in AI, digital transformation or product, at a company building what comes next: tech, AI, space, semiconductors or biotech. Based in Valencia: remote or hybrid preferred, open to relocation.

Languages

Spanish native
Catalan native
English C1 (Cambridge)
French basic
German basic
Mandarin Chinese learning, basic

Outside work

Urban sketching
Fitness
Strategy games
LOTR geek
Neuroscience
Financial markets

Selected work

AI and optimisation in operations

Three tools I built at DHL. Their code and data are confidential, so two demos replicate the patterns with synthetic data and a generic design.

AI agents · forecasting · data quality

Airbus Forecasting Tool

DHL Supply Chain · Airbus account · delivered 2026

Problem: Forecasting annual spend against budget was a fully manual process across several Excel files, with serious forecast errors.
What I built: A Python tool built with AI agents that automates the cost forecast, data-quality validation, the deep dive (a year-over-year cost bridge down to route and accessorial cost) and monthly reporting. 26 automated checks (13 blocking, 13 warnings) gate the data, and a person must validate the forecast before it is published.
Built with: Python and AI agents, inside DHL's corporate Microsoft environment (Copilot Studio, MCP connections to SharePoint), so it runs within the company's systems rather than as a standalone tool.
Result: It serves 18+ Airbus transport managers and the DHL account team, and saves the equivalent of a senior controller every month. Decisions are already made with it: the budget was raised by 15% before year-end, based on a more detailed forecast that includes market disruptions such as fuel.
What it taught me: Success metrics need to be explicit. I defined them for this tool, targeting cost data quality from 78% to 98%, year-end forecast accuracy from 75% to 95%, and the full cycle from one month of a senior controller to 2 days targeted

Try the synthetic-data demo

Multi-agent system · tenders · data quality

RFQ Value Creation Engine

DHL Supply Chain · new-client tenders · first live tender July 2026

Problem: To win a tender, DHL has to show a new client how much it will save them, from the client's raw data and in 1 to 2 weeks. The saving has to be built up from the client's own transactions, lever by lever, because an inflated number turns into a commitment the client will hold us to.
What I built: A multi-agent AI engine that applies DHL's Value Creation levers bottom-up: an orchestrator and 11 specialised agents (client context, data gaps, data quality, baseline, lever screening, sizing, benchmark calibration, projection, traceability, deck, and a red-team challenger), with 4 expert gates where I decide: the baseline, the lever shortlist, saving % and confidence, and the final sign-off. The LLMs read, classify and write; every figure is computed by versioned Python with reconciliation checks.
Built with: Python and LLM agents, used on a live tender; being rebuilt on DHL's Microsoft Copilot stack.
Result: On its first live tender, for an automotive OEM, it turned around 20K shipments into a €2.05M three-year savings programme projected in about two working days instead of 2 to 3 weeks, traceable shipment by shipment and delivered with the deck and a live 3-year model.
What it taught me: Reconcile before you analyse, and let the data overrule you: it demoted the team's top hypothesis and turned round trips, which nobody had raised, into the main finding. Agents accelerate; the expert decides.

Try the synthetic-data demo

Optimisation · network design

Road Network Optimizer

DHL Supply Chain · Airbus Defence & Space · delivered 2025

Problem: Finding route configurations better than the current ones, using the rates the 3PL pool offers today or could offer.
What I built: A Python algorithm that finds alternative route configurations and optimises cost, lead time, CO2 or a combination of the three. Applied to around 100K Airbus Defence & Space shipments a year.
Result: It generated a project initiative sold to the client to optimise road routes between plants and suppliers in Spain, with around €400K expected annual savings projected. I presented it to Airbus and DHL leadership.
What it taught me: Showing the work matters. Presenting this tool and the Forecasting Tool to Airbus and DHL leadership increased the client's confidence in DHL's digital capabilities.

Related demo: Tender Savings Engine No demo of its own; the Tender Savings Engine routes and optimises a related network on synthetic data.

Try it

Demos

Synthetic-data replicas of tools I built in production.

Demo 1 · forecasting · validation · AI agent

Cost Forecaster

The monthly cost forecast, end to end: it runs in seconds, validates itself, explains every euro of change and waits for a person's GO, with an AI agent briefing the reviewer.

Tested on 39 reference cases: 100% of planted data errors caught, 98.5% year-end accuracy.

Open Cost Forecaster

Demo 2 · tenders · optimisation · AI agent

Tender Savings Engine

From a client's raw shipment file to a bottom-up savings plan: a data check, a solver that re-plans a European road network, levers you can calibrate, and an AI-drafted proposal.

Tested on 30 synthetic clients: 100% of planted data errors caught and of the planted saving found.

Open Tender Savings Engine

What I bring

Skills

Supply chain and operations

Network design and route optimisationHub and spoke, cross-docking, milk runs, linehaul, modal shift (short-sea and rail), for aerospace and consumer goods clients
- Log-hub
- Lyric (Coupa)
- Python
Savings strategy and cost controlAnnual savings strategy, year-over-year cost bridges, savings validated against projections
Capacity and last-mile planningCentralised planning at 1 week, 48 h and 24 h for 40 delivery stations
Tender due diligenceExpress 5-year savings analyses for new clients in energy, automotive and pharma

AI and data

AI product managementSuccess metrics, data-quality gates and mandatory human validation; Duke AI Product Management
AI agents in a corporate stackAgents in Python; Copilot Studio and GitHub Copilot; MCP connections to SharePoint and OTM
- Python
- Copilot Studio
- GitHub Copilot
- MCP
ForecastingCost forecasting with AI agents; ML forecasting of shipment volumes
Data and reportingPython, SQL and VBA; cost-control reporting designed end to end, with Power BI implemented by the Data & BI team

Leadership and change

Team leadership4 Value Creation Managers in Spain and France at DHL; 8 logistics analysts at Amazon, with hiring authority
Executive governanceWBR, MBR and SBR, root-cause analyses, and deviations presented to Airbus and DHL C-suite
Change managementNew contractual KPI package rolled out against strong internal resistance; led the move to an AI-native team

Career

Experience

Senior Value Creation Manager

Jan 2024 – present DHL Supply Chain · 4PL, Lead Logistics Partner for Airbus Defence & Space and Commercial

Lead 4 Value Creation Managers in Spain and France, each owning a scope of the logistics network as an autonomous project manager. Main contact for Airbus and DHL leadership; run WBR/MBR/SBR governance, RCAs and action tracking.
Own the account's annual savings strategy and validate real savings against projections. Around €2M of savings over 3 years are attributed to me personally.
After the 2025 contract renewal, designed and rolled out the new contractual KPI package against strong internal resistance, in time to turn the KPIs green and keep the business.
Built an AI-agent forecasting system that replaced a fully manual, spreadsheet-based cost forecast for Airbus, with 26 automated quality checks and mandatory human validation; used by 18+ Airbus transport managers.
Built a multi-agent AI engine that sizes savings for new-client tenders; on its first live use, an automotive tender, it produced a €2.05M three-year savings programme (projected) in about two working days, instead of 2 to 3 weeks.
Led the move to an AI-native Value Creation team and trained colleagues on Copilot Studio, GitHub Copilot and MCP connections to SharePoint and OTM (Oracle Transport Management).

Operations Manager I

Jun 2022 – Dec 2023 Amazon Logistics · EU Central Operations

Led 8 logistics analysts running centralised labour capacity planning for 40 delivery stations in UK-MEU.
Designed and rolled out a simplified last-mile capacity planning process run from the Barcelona control tower: projected 3.2MM savings in year one and around 5-7% more average daily capacity. Live in 40 UK stations, with the rollout started in around 30 in Germany and Poland.
Built Python planning systems (1 week, 48 h and 24 h ahead) that improved planning accuracy by around 25%.

Logistics Analyst, Team Leader

Oct 2020 – May 2022 Amazon Logistics · EU Central Operations

Simplified last-mile delivery scheduling across EMEA and led process analyses with SQL, VBA and Python.

Supply Chain Engineer

Jun 2019 – Oct 2020 SRG Global · automotive parts

Led supply chain improvement projects with an average 25% logistics cost reduction and a payback under 1.2 years, and plant and warehouse layout redesigns covering over 100,000 m².

Machine Learning Engineer

Jan 2019 – Jun 2019 PFS Tech · IT consultancy

Built and deployed an ML and RPA invoice-processing system for a large retail client, valued at USD 100K.

Education

MSc Strategic Operations and Innovation Management
OBS Business School – Universitat de Barcelona, 2024
MSc Industrial Management Engineering
Universitat Politècnica de València, 2019
BSc Industrial Engineering
Universitat Politècnica de València, 2017

Certifications

AI Product Management
Duke University
Google Project Management Professional Certificate
Microsoft Certified: AI Agent Builder Associate
Candidate, exam in November 2026

Side projects

Swing Desk (SWINTR)

Live, for my own trading, since October 2026

My trading desk on the phone. It watches my portfolio and the plan behind each position, sends Telegram alerts when a rule is close or triggered, keeps a decision journal, and every night delivers a briefing that replaces scrolling X: news, portfolio, watchlist and market regime.

AI agents triage the news against each position's thesis and write the nightly briefing; a rules engine decides and code checks every number. Next.js app, Supabase, Python jobs on GitHub Actions, Gemini and Claude, under $1 a month in LLM costs.

Investment Briefing

In production since May 2026, real subscribers

A daily investment newsletter written end to end by an orchestrated team of AI agents: a market-regime analyst, a Substack summariser, an idea generator, a validator and an editor that assembles the email.

Runs every day on GitHub Actions. Agents on Gemini Flash, connected by API and MCP to Substack, macro data and market prices.

Contact

Get in touch

Email or LinkedIn.

facundobertuajobs@gmail.com Open in mail app

linkedin.com/in/facundobertua