Ximena Quintana

I turn messy operations into systems that scale.

Bring me the messy stuff.

Twelve years inside mobility, logistics and marketplaces. Now building AI-assisted operations that a team can actually run.

Your ERP is late, the process lives in three spreadsheets, and nobody agrees on the workflow. That is the job.

See my work Download CV
Start a project How I work
Twelve years across DaimlerRyderStellantisMercado LibreUrbvanKolorsUberOpenRide

I speak operations. I speak tech. I translate between them.

The interesting problems live right there, in the middle.

  • Ops

    Operations

    Fleet and supply operations, dispatch, scheduling, route performance, SLAs, capacity, control tower.

  • Sys

    Systems

    Odoo, Asana, ERP implementation, integrations, data structures, geospatial, adoption.

  • AI

    AI

    Automation, agents, copilots for process discovery, knowledge systems, internal tools.

Right now

I build the tools I use.

I consult on operations and implement management systems shaped to the client. Then I build the tooling that makes that work, including the MCP servers that let an AI assistant actually operate a platform instead of describing it.

  • Own OpenRideA mobility operating system for taxi fleets, airport transport and B2B ground transportation. Mine, from zero.
  • MCP Servers for CloudTalk, Asana, ClickUp, Miro and PipedriveThe CloudTalk one exposes 86 capabilities: 70 tools, 9 resources and 7 prompts.
  • Scaffold Scaffolding for building apps and skillsTooling for building tooling, so the next one starts further along than the last.
  • Client Management systems on Odoo and Asana, shaped to the operation

Trajectory

Twelve years of operational puzzles.

2025 OpenRide Founder and product lead. A mobility operating system for taxi fleets, airport transport and B2B ground transportation.
2025 Xmarts Group AI systems and product consultant. Eighteen client implementations across ten industries, plus the internal tooling and MCP servers behind them.
2024 Uber Supply manager, Uber Shuttle. Built the high capacity vehicle supply behind new urban shuttle operations in Mexico City.
2022 Kolors and Urbvan Product operations and scheduling lead. Transport inventory, routing and the control tower view of a shared mobility platform.
2021 Mercado Libre Senior logistics engineer. Process design and automation inside the largest marketplace fulfilment network in Latin America.
2017 Ryder and Stellantis Logistics engineer for automotive. Network design, launch logistics and cost per unit for FCA. Also the supplier debt nobody owned, the CFDI billing and addendas, and the transport requests the system could not pay by itself.
2014 Daimler, BASF, Gepp Where it started. Spare parts and warranty analysis, material planning, and a route control and fuel optimisation project.

Selected work

Four problems, and what I did.

Most of this lives inside other companies and I cannot show the artifacts. So here is the shape of the problem, what I built, and what changed. The pattern repeats more than the industry does.

Fleet partner onboarding system

2→15 fleet partners in the Uber Shuttle supply base
Challenge

A global mobility platform needed local fleet partners in Mexico, had no repeatable way to onboard them, and did not understand how those operators actually work.

What I did

Built onboarding from the operator's point of view: clear requirements, a documented flow, SLAs both sides could hold, and a feedback channel back into product.

Result

The network went from a pair of partners to a working supply base of fifteen, and the platform changed how it thought about local ground transport operators.

The debt nobody owned

Challenge

Millions of pesos of supplier debt sitting unresolved between accounts payable and the suppliers themselves. It was not in anybody's job description, so it was nobody's problem, and it kept growing.

What I did

Took it without being asked. Connected accounts payable directly with the suppliers, worked the backlog case by case, and held the control myself until it was clean.

Result

The backlog came down to a fraction of what it was. The part that lasted longer: there was now one person who could say what was owed, to whom, and why.

What can legally ride together

Challenge

Freight planning had to combine hazardous materials in the same load. Whether two classes can share a trailer was a judgement call that lived with whoever had been there longest. Getting it wrong is a safety and compliance problem, not a delay.

What I did

Built a compatibility matrix that takes the hazmat classification of everything going into a load and returns what can travel together and what cannot. The judgement became a rule anyone could apply.

Result

Planners stopped queueing behind the one person who knew. The answer moved into building the load instead of being a check after it.

Live Rebuilt it. Load an order and watch it split.

Relay points across the border

Challenge

Cross border freight loses its time at the crossing, and the whole trip waits on it. That time is paid by drivers, by equipment sitting still, and by the customer's delivery window.

What I did

Designed a relay point system. Instead of one driver taking a load the whole way, the route breaks at a relay, so the crossing becomes one designed stage rather than the bottleneck of the entire trip.

Result

The network stopped absorbing the crossing as dead time and started planning around it, with equipment and drivers where the route actually needed them.

It keeps being the same move.

  • Cost per unit, before committing

    Automotive logistics is priced per unit and budgeted a year out. I ran the logistics studies and the cost per unit analysis for new product launches and annual budgeting, and became the person who knew the client's internal logistics system well enough to get real numbers out of it, in Excel and Power BI.

  • Improvements in the tools they already had

    On site with a global electronics manufacturer's operation, improving how it ran using Excel Online and Power BI. Not a new system nobody would adopt. The tools the team already opened every morning, doing more.

  • Capacity that had to exist by Tuesday

    Marketplace demand does not wait for permanent capacity. I designed a temporary fulfilment facility for one of the largest logistics networks in Latin America: layout, flow, and the operating assumptions it had to hold.

  • Scheduling that stopped living in one head

    Shared mobility scheduling had to balance real vehicles, real drivers and passengers who do not care about the constraints underneath. I turned transport inventory, scheduling and routing into a system with rules: how capacity is assigned, how routes are built, how exceptions are handled.

  • Whether a route pays for itself

    Intercity routes have to earn their cost, and whether a given one does is not obvious until somebody works it out route by route. I ran that costing, so the decision to open, hold or close a route stopped being a feeling.

  • The invoicing change of 2018

    I closed the development of the billing EDIs and the addendas. In automotive it is the customer's required format, not the tax rule, that decides whether an invoice gets paid, and every customer wants a different one.

  • I became the exception path

    The system paid the standard moves by itself and left the special transportation requests to a human. I became the specialist in those, which is a polite way of saying I became the path the system did not have.

  • A note on the tools

    Some of this was built in Excel, because Excel was what the operation had. I would build most of it differently today, and I do. The tool was never the point: the point was turning a judgement that lived in someone's head into a rule the operation could run without them.

What I can help with

Four ways this usually starts.

Process design

Operations and process design

When the operation grew faster than its processes. Map what is actually happening, find the bottlenecks, redesign the workflow, define ownership and KPIs.

Implementation

ERP and software implementation

You bought Odoo, Asana or something else, and it still has to become a real operation. Requirements, process design, migration, integrations, adoption.

Automation

AI automation

Not AI because AI. Find the repetitive work humans should not be doing any more, and build the agent, the automation or the internal tool that takes it.

Rescue

Operations rescue

The implementation is stuck. The process does not match the software. Nobody agrees on what the workflow should be. This is the one I like most.

  • Our ERP implementation is six months late.
  • Everyone uses a different spreadsheet.
  • We bought the software but nobody knows how it should actually work.
  • Our operations team spends half the week copying data.
  • Engineering built what we asked for, and it is not what operations needed.
  • We know AI could automate this. We just do not know where to start.

Method

This is basically what I do.

Mess WhatsApp Excel ERP Email approvals Tribal knowledge AI tools

  1. Understand the operation

    I go where the work happens and learn how it runs, including the shortcuts nobody documented. You cannot fix what you have only seen in a diagram.

  2. Map it and find the leaks

    What I saw becomes a clear map of the process, with the exact points where time, money and quality leak out. Priorities become obvious here.

  3. Build with AI, review as a human

    I move fast using AI to build the copilots, the documentation and the automations, then I review every output myself. Speed without judgement ships mistakes.

  4. Hand over something they can run

    Documentation, onboarding, and a team that knows how to keep it going. My work is done when the operation runs well without me in the room.

How I work

I do not just learn tools. I reverse engineer how they are meant to work.

Give me a tool I need to use and I will end up figuring out the whole thing. I keep becoming the unofficial super user of internal software, not because it was assigned to me, but because I keep asking what else it can do. That is also why implementation goes well: I am not looking at buttons and modules, I am looking for where the logic of the system and the logic of the business are fighting each other.

Three times, so far

At Stellantis my title said logistics engineer. It said nothing about their internal logistics system. I needed it, so I started poking at it, and ended up as the person who understood it well enough to get real numbers out of it.

At Kolors it happened again with their internal route management system, and that time it turned into the job: I became the person who translated operations into product, because I was the one who could speak both.

It has happened with every system since.

And the unglamorous part

When Kolors launched its US pilot I helped run the operation and took the user calls myself, on top of everything I already owned. You hear things on those calls that never reach a dashboard.

On the side

I also translate nerdy things.

Explaining a system to someone who does not work in tech is the same muscle as running an implementation. If the people who have to use it do not understand it, it is not built.

Traduciendo Nerdadas

Making AI and technology understandable for normal humans, in Spanish.

Ximena Quintana

About

Industrial engineer by training. Operations person by instinct.

I have spent my career inside companies where technology, operations and people have to somehow cooperate. I like figuring out why something is not working, mapping the chaos, and turning it into a system people can actually use.

I grew up and built my career in Mexico City, and I work from the San Francisco Bay Area now. Both cities are full of jacarandas, which is where the purple on this page comes from.

San Francisco Bay Area Industrial engineering, Universidad La Salle Spanish and English Twelve years in operations

Contact

Tell me what you are building.

Got an operational mess?

If you are hiring for operations, product operations or applied AI, I would like to hear about the team and the problem.

You do not need to know which software you need, and you do not need the project defined. That part is my job.

Role, team, and what the first ninety days would actually be about.
Where are you
This opens your email app with the message ready.

Your email app should have opened.

If it did not, here is your message. Copy it and send it to the address below. Nothing is lost.

ximena@ximenaquintana.com

Or write to me directly at ximena@ximenaquintana.com