Libra Investimentos
Dashboard design and development for an investment office to help centralize and visualize financial data from client's wallets.
The Libra's investment advisors team and the management team needed a way to visualize their clients data in a more concise and organized way, helping them to be quicker and more precise on their meetings, alignment between teams and decision making.
This data was distributed across different spread sheets, these were fed by data retrieval automations from other sources. The data involved clients wallets, investment assets, values attributed to each investments, sales commission, ROA (Return Over Advisor), Rentability, Clients Buy/Sell Strategies operating on 3rd party Trading Robot services,
Along side with the teams, I decided to build a dashboard to unify all these different sources and provide an interface that prioritized both Advisor's and Management's teams needs.
I started the design process interviewing the teams leaders and designed a low-fidelity prototype in Figma, so that we could align if the right information was being placed where it should. After rounds of iteration, I finalized a high-fidelity prototype, which served as a basis for the development of the applications Front-End.
Taking it from prototype to working product was where I stepped outside my usual scope: I built it with Claude Code, which made it realistic to reach a functional MVP rather than stop at a clickable prototype.
Next.js with Tailwind and shadcn/ui kept the visual decisions from the prototype under my control through implementation, working around a legacy database the desk already used. It was hosted in Vercel in a private link as the team started testing it out on their daily workflows.
This is MVP-grade code that did its job as an internal tool, and it would go through an engineering review before reaching full production.
Interviews with the squad leaders shaped what the dashboard shows and, just as importantly, what it leaves out; the filters follow the questions people actually asked rather than the shape of the data.
Two features came only after the MVP was already live: notifications over the active strategies, since expiration dates were being tracked by memory, and client observations, so decisions get recorded next to the data instead of somewhere else.