Moni

Where they were

Moni is an Argentine lending fintech. Before the pandemic they had a solid data team that managed the very heart of the business: the risk model. However, after the pandemic, that team started falling apart: people left for other companies, and with each departure a piece of critical know-how was lost. It wasn’t so much a sudden crisis as a slow erosion that ended up very strongly affecting the quality of operations.

What we found

The initial request was help with the risk model. But the real problem was deeper: the data was chaotic, the ETLs were pure legacy, and there was no base to build anything reliable on. The most valuable thing we did wasn’t a technical deliverable. It was the diagnosis: telling the team exactly how far they were from where they needed to be.

What we built

A delinquency prediction model that laid the foundation for Moni’s team to build their own models with total autonomy. Complete migration of legacy ETLs to modern architecture: with these changes, what used to take weeks in pipeline development reduced to one-fifth of the time and a consequent 50% drop in operational costs. We rebuilt the model training infrastructure that tripled the speed of time-to-prod. But the most important thing is that the work was complemented by solid knowledge transfer not only to the technical team, but also to the business team. Open, fluid, and honest communication between Moni and deployr allowed us to develop solutions proportional to the fintech’s real size, without over-engineering.

What they can do now

Moni’s team builds their own models. Today they have a new customer model that controls delinquency in a tough macroeconomic context (high default rates in Argentina since late 2024) and it works. Recovery in the high-risk segment improved 20%. The knowledge isn’t locked in anyone’s head: it lives in the infrastructure and processes. If the team changed completely tomorrow, Moni wouldn’t start from scratch.

Why they came back

19 months. The relationship extended because the value kept showing up. After a previous experience with a more corporate provider (closed scope, not going an inch beyond what was agreed) the difference was clear: personal investment, founders present, honesty even when the truth was uncomfortable.