Pablo G.

You're probably generating more data than you know what to do with. Most of it sits in dashboards nobody opens, or feeds reports nobody trusts. I fix that.

25 years building data infrastructure for companies from startups to Fortune 500, across the US, Latin America, and Europe. On Upwork since 2016, Expert-Vetted (their top 1% of independent talent). I'm not an agency, and I don't do offshore handoffs. It's just me, doing the work you need done. And my work history can testify that I can get it done.

I've been the fractional CTO reporting to CEOs at data-heavy companies, the data lead serving a C-suite, and the extra pair of senior hands inside an existing team. The structure can be flexible, but I'll never loosen the standards.

What I actually deliver:

Software that ships. When the fix is a tool, I build the tool and put it in front of users. A quote builder that a client's sales team runs every day. A multi-tenant data platform I designed, built and operate myself on a daily basis. Agents that write back to your systems instead of just reporting on them. I'll talk architecture with your engineers when it matters and keep it out of the room when it doesn't. I hate prototypes that never leave the sandbox; if it's useful, it's in production and it's running.

Reporting people trust. Not dashboards that look impressive in meetings, but numbers that trace back to source and survive hard questions. A recent client discovered they'd been overstating ARR for years. I rebuilt it from transaction data and now the figure holds up in due diligence.

AI and ML that reaches production. I've shipped forecasting models, document extraction, and analytics agents that run daily in live businesses. One recent model beat the client's specialist analyst on 7 of 10 product categories. Where it lost, I told them why, because knowing the limits is part of the job.

Audits that find real money. Pricing analysis for a SaaS client surfaced $193K a year in unintended cross-subsidies between customer tiers. A BI audit for a wholesaler found board reports that couldn't be traced to source, plus the path to fix them.

Integrations that stop being fragile. QuickBooks, Unleashed, Salesforce, SAP, Chargebee, Snowflake, and the long tail of tools mid-market companies actually run.

Stack: Python, R, SQL (PostgreSQL, MySQL, MS SQL), Snowflake, AWS, Azure. BI: Qlik Sense, Power BI, Tableau. AI/ML: Anthropic Claude, OpenAI, custom models. Plus GIS, forecasting, and marketing analytics when the problem calls for it.