I'm Jones Madruga — Fractional CTO and AI advisor. Twenty years shipping production AI for fintechs, scale-ups, and the boards that fund them. Hourly access to senior judgment, no retainers.
Most engagements anchor on one or two of these. The same framework whether you're a Series A founder figuring out where AI fits, or a leadership team that wants an experienced outside operator helping shape the roadmap.
How your engineering team uses AI to ship faster, with higher quality, at lower cost. Workflow integration, agentic tooling, governance, developer adoption. Not Copilot evangelism — operating model design.
How AI creates real value for your customers — not as a feature checkbox, but as a wedge. Roadmap, architecture, build-vs-buy, and the hard tradeoffs between speed, cost, and reliability when you're shipping AI at scale.
How product, ops, marketing, finance, and legal teams adopt AI day-to-day. Different solutions, equal importance — the compounding gains usually live here, and most companies are leaving them on the table.
Selected results from operating roles. The same instincts and frameworks I bring to advisory engagements.
Designed and shipped end-to-end. Customer-grade quality at scale across millions of conversations.
Customers engaged on complex renegotiations in a safe space, on their own schedule, with personalized offers — end-to-end automation, no human handoffs.
Internal agentic tool reduced policy and test validation cycles from ~2 weeks to ~4 days, directly improving engineering velocity.
Architected and shipped an in-app IPO investing flow inside an app used by 48M+ users.
~R$600K/month savings, availability up from <80% to 98.7%. Engineering turnover dropped from 22% to 8.9%.
From ~800ms to ~7ms; P99 from ~5s to ~95ms. ~97% reduction in DynamoDB rules-table cost with zero feature impact.
I've spent 20+ years building and scaling customer-facing platforms across fintech, edtech, and mobility — combining deep backend, distributed systems, and AI/ML expertise with a record of hands-on technical leadership.
Most recently I was Director of Engineering at Nubank (NYSE: NU), leading the Global Collections platform — serving 50M+ customers and enabling ~$6B in cash recoveries — where I shipped AI-integrated experiences across WhatsApp and Voice channels and drove ~99% latency reductions in production.
I've held four CTO seats — at Arco Educação (NASDAQ), Tembici, Stoodi, and Immplay — built data platforms processing terabytes of unstructured government data, and teach applied Machine Learning at Sirius.
Jones helped us move from AI curiosity to AI execution. He understood our product, our operating constraints, and our team maturity, then turned that into a practical implementation roadmap. Within weeks, we had clearer priorities, faster internal adoption, and a much better sense of where AI could create real leverage.
Rare combination of deep engineering judgment, platform thinking, and pragmatic AI implementation experience. He helped us think through architecture, governance, developer workflows, quality controls, and adoption at scale. His work raised the technical bar while keeping the team focused on measurable impact.
An exceptional partner to our portfolio companies on AI implementation. He helps founders avoid the two common traps: overbuilding AI infrastructure too early, or applying AI only as a superficial feature. Clear path from strategy to execution, with the right balance of speed and commercial relevance.
Three formats. Most clients start with hourly access — no retainers, no minimums, no awkward off-ramps.
Senior judgment, hour by hour. Most clients use 8–24 hours per month. The fastest way to start, and the easiest to stop.
For specific work — an AI strategy, an architecture review, a CTO assessment, a board-ready roadmap. Fixed scope, fixed price, clear deliverable.
Independent board seats and formal advisor roles for companies where I can be useful for years, not weeks.
A 30-minute first call is the cheapest way to find out if we should work together.