Eduardo Rueda · Dubai, UAE
Business Process Analyst — Process Mapping & Workflow Automation. On-site · Hybrid · Remote.
Who
I map, redesign, and automate business processes — turning operational complexity into systems that run cleanly and scale. At MedStar Diagnostics, I inherited a fully manual pipeline: patient intake, insurance routing, billing, clinical report formatting, and reconciliation, all done by hand. Over 6 months I mapped every step, found where time and accuracy were being lost, and built 12+ automated workflows — using Python, Excel, and Claude AI — that now run the pipeline end-to-end. Processing time dropped 60–75%, and the system was adopted as company standard, not because I was asked to build it, but because it worked.
At TOTVS, Latin America's largest ERP provider, I spent 3 years as an ERP Consultant leading implementation projects across manufacturing, supply chain, procurement, and cost management, for clients in Colombia and Ecuador. That's where I learned to read a process from the inside: sit with the people doing the work, find where the system breaks, and rebuild it so it holds.
Why I do this: I don't just want to make a process faster. I want to leave behind something the team can run without me — a system, not a dependency.
Why
I feel like a craftsman who is called every time someone needs a solution or tool to help others find what they can do.
Not only do I transform my own life through my work — I create value out of it, and bring that to others through win-win partnerships.
Technology isn't the point for me; it's the gear I use to move things toward meaningful paths. The first time I used AI to clear an inbox that had been accumulating for over ten years, what I found wasn't an empty inbox. It was a version of myself that was organized, systemic and capable. The technology didn't create that version. It cleared the way to it.
The same clearing-the-way pattern is how the MedStar system actually got built — not as a plan, but one bottleneck at a time. I started by running the whole process by hand, one exam and one report at a time. Claude removed the tedium first, then I found Skills, then Projects. Eventually I hit a real ceiling: I was burning through my own token budget on a subscription I paid for myself, and waiting for a reset wasn't how I wanted to work. So I built a version that didn't need Claude at runtime at all. The tool didn't just save time — it removed a dependency I'd created for myself.
The same thing is true of organizations. A good system doesn't just save hours. It clears the way to the team that was already there — the one that was buried under manual work and couldn't be seen.
What
I learn a process deeply, then I build systems that make that process scalable and owned by the organization.
Every workflow I've automated, I ran by hand before I touched it. You cannot automate a process you only understand from a diagram — the exceptions are the process.
Not everything should be automated. I look for the steps that are repetitive, rule-based and verifiable, and I leave judgment where judgment belongs.
A system that works on clean sample files is not a system. It has to survive the typo in the folder name and the record that doesn't match.
The measure of success is that it runs without me and someone else can maintain it. I document while I build, not afterwards.
AI & automation
Data & development
Operations & delivery
How it helps
If one of these sounds like your operation, we should talk.
Volume grows, the work is manual, and the only lever anyone sees is hiring. I find the repetitive core, automate it, and give you capacity you already paid for. At MedStar, processing time dropped 60–75% — and the gain gets larger as patient volume rises.
Tribal knowledge is a risk you don't notice until someone leaves. I turn undocumented routines into documented, auditable systems — and I write the manual as I build, not in the last week of the project.
I don't install ERP — I help organizations adopt it. Four years at Latin America's largest ERP provider — three of them leading end-to-end implementations — taught me that the technical go-live is the easy half. The work that decides the outcome is training, testing and sitting with the people who have to use it.
Clinical, financial and personal data often can't go near a cloud API. I've built a full automation pipeline that runs 100% offline with no external calls, and a human approval gate before anything is finalized. Constraints like this are a design input, not a blocker.
Selected work
Each case study includes what I'd do differently. Defensible beats impressive.
Healthcare · 2025–2026
12+ interconnected workflows covering patient intake, clinical report processing, billing reconciliation and delivery — built as a personal initiative, adopted as company standard.
60–75% processing time reduction
Healthcare · 2026
A tracking system for an operation where three patient cohorts are always in flight at once, built to survive the messiness of real archive data rather than assume it away.
One front door for daily operations
ERP · Manufacturing
A ransomware attack wiped a quarter of a client's inventory and billing records. I recovered them from an untouched BI server with MAQL, SQL, and ADVPL.
100% of records recovered
Earlier results
Contact
I'm open to roles in Dubai and Abu Dhabi, and to fully remote positions internationally. The fastest way to reach me is email.
