Selected work
Operations we run, at the scale we run them.
Three operations that show the arc of how we work: the problem an operator brought us, what we built, and what is live in production today.
- Enterprise AI OS · multi-region
Agent control layer for finance ops
The problem. A finance team had a dozen useful agents — close, reconciliation, vendor onboarding — running as scripts no one outside engineering could see, audit, or trust.
What is live. Enterprise AI OS gave operators a governed surface for every run: roles, owners, KPIs, and replay. Controllers approve from one queue. Daily close lands on time, and 'what did the agent do?' has a one-click answer.
15+agent workflows now governed and live - Payments · Dominican Republic
PagaTodo · Grupo Vimenca acquiring
The problem. One of the country's largest acquirers needed switching, terminal apps, and merchant tooling that could carry national-scale traffic without a vendor cycle dictating the roadmap.
What is live. Seven years on the bridge with the same engineers who designed the system. Settlement, disputes, and approvals run on infrastructure we operate alongside the bank — not behind a support desk.
2M+payment transactions processed monthly - Retail SaaS · Chile
Operating platform — multi-brand retail
The problem. Skechers, Tecnobuy, and other multi-brand groups were running back-office, marketplaces, and store operations across spreadsheets, integrations no one owned, and weekly fire drills.
What is live. One platform for catalog, pricing, orders, and store ops. Marketplace listings stay healthy across seven channels. The same team that built it is still on-call — going on year seven.
7 yrsthe same engineers, seven years later