Applied AI lab
Here I publish what I build with AI and data for compensation and People Analytics. Every project is live and you can open it.
By Federico Gómez · People Analytics and compensation · Bogotá
The connector is the only person with ties in all three areas. If she leaves, Operations, Sales and Technology have no ties to each other.
Projects
Turnover, collaboration networks and salary bands, each in its own tool.
Executive overview, turnover detail and an absenteeism and talent view, with the narrative ready for a committee. It shows why good talent leaves and what it costs.
A collaboration network you can explore: who connects the areas and where the silos are. Pick a person and the network shows what happens if they leave.
It places each person against their band and the market, and shows which salary zone they are in. Bands by curve and grade, market percentiles and indicators. It is the lab’s first project designed as a product.
Interactive examples
Hover over or tap the charts to see the detail.
Each dot is a person, placed by salary relative to their band midpoint (compa-ratio). The shaded strip is the band: 0.8 to 1.2 times the midpoint.
Monthly absenteeism rate by area in 2025, from the talent dashboard data. With the pattern in view, the peaks can be planned for.
Notes
Pieces on agentic AI, organizational networks, compensation and People Analytics, with examples.
Data and limits first, then the agent: the order I suggest.
Who connects teams, where the silos are and what is lost when someone leaves.
How to read the compa-ratio, your position against the market and the out-of-band share.
How it gets built
I measure the problem in hours and money: which process takes the most time, which decision is made without data and what data exists.
I put the architecture, the scope and the success criteria in writing before building.
I build in short deliveries, each one ready to test.
I document the decisions and how it runs, so someone else can maintain it without depending on me.
This is how I build an agent: it works on governed data and checks each step before it delivers a result.
The lab
SintéticaLab is my applied AI and analytics lab, based in Bogotá. Many HR teams already have payroll, performance reviews and surveys, but each source lives on its own. Here I test how to connect them to answer concrete questions: who is leaving, what it costs and who you can’t afford to lose.
What I publish here comes from working with corporate data platforms (Microsoft Fabric, Azure) and with analytics and development teams.
Today the lab works as an open portfolio: I explore a problem, build a complete solution and publish it. Some projects, such as the salary band calculator, may become products.
Stack
Who is behind it
I work in People Analytics, compensation and building data and AI solutions. SintéticaLab is the lab where I build and publish what I try out.
Each project is published once it works end to end.
If a 200-line script solves the problem, there is no need for an agent.
I check every figure against the source data until it matches 100%.
If something here is useful or you have a similar problem on your hands, get in touch.
I reply in Spanish or English.
If you write, I use your details only to reply. Data policy.