2
Active clients in the construction sector
At Vidra, we combine data science, software engineering, and strategic consulting to create useful, measurable, and easy-to-adopt solutions.
We work with companies that want to apply artificial intelligence with rigor, clarity, and focus on real results.
3
Projects in production or in testing
9
People in the team
2
Project types: ML inference and agentic flows
Team
The people who do the work.
We combine expertise in data science, software engineering, and project management to deliver AI systems that run in production.
JM
Jose Antonio Morales
Partner
JB
Jorge Bernal
Partner
JS
Juan Sánchez
Partner
IN
Ignacio Navoz
Technical lead · Governance officer
JC
Javier Cerón
Infrastructure Engineer · Cloud Architect
BSc in Applied Mathematics · MSc in Artificial Intelligence
AC
Álvaro Carrera
ML / Full Stack Engineer · DevSecOps Engineer
BSc in Computer Engineering (Software Engineering)
CD
Carlos Díaz
Frontend Engineer · LLM/Prompting Specialist
BSc in Applied Mathematics · MSc in Artificial Intelligence
GH
Gonzalo Humanes
Data Analyst · Data Engineer
BSc in Applied Mathematics
What makes us different?
Technical clarity, business judgment, and responsible execution.
01
01 / Absolute Transparency
We explain the why and how behind every algorithm. We design white-box models where the client owns and understands the core logic of their data.
02
02 / Engineering Rigor
AI is only as good as the data that feeds it. We prioritize clean data architecture and impeccable pipeline engineering before training any predictive model.
03
03 / ROI-Driven Approach
We do not build technology just for the sake of innovation. Every project at Vidra is designed with clear commercial KPIs agreed from day one.
Talent
We bring technical, strategic, and creative perspectives together to solve problems from multiple angles.
Dedication
We care about delivery, detail, and the real adoption of each solution.
Impartiality
We recommend what the business needs, not what sounds most impressive in a demo.
Our way of working
Principles that hold every project together.
What we build
Three types of system already in production for our clients.
Predictive ML models
Construction cost estimation with LSTM + CatBoost + LightGBM, trained on real historical data and deployed as APIs with confidence intervals.
Agentic pipelines
Automated weekly detection of private tender opportunities via multi-agent analysis with Gemini AI, delivering scored results with 0–100 ranking and commercial feedback.
Modular AI platforms
Extensible hubs (React + Vite + FastAPI + Docker + Electron) organized by department. Adding a new feature means editing a configuration registry, not a rewrite.
01
We understand the context, goals and real constraints of the business.
02
We identify opportunities where AI can create measurable value.
03
We design practical, scalable solutions aligned with existing teams.
04
We support execution so transformation reaches day-to-day operations.
Our process
Designed to turn technical complexity into operational progress.
Our people
A team of rigorous experts, focused on delivering operational and high-impact AI solutions.
Our team combines diverse expertise and a collaborative culture, focused on maximizing the performance and applicability of every AI solution for our clients.
Explore careersML Inference
LSTM, CatBoost, and LightGBM models deployed as synchronous APIs with confidence intervals, trained on real historical client data.
Agentic flows
Asynchronous multi-agent pipelines with Gemini AI for automated detection, analysis, and scoring of market information.
RAG & chatbots
Chatbots over a client's own documentation using pgvector, hybrid vector + lexical search, and OpenAI.
Modular platforms
Extensible React + Vite + FastAPI + Docker + Electron hubs organised by department, built to grow without rewrites.
Companies we have worked with
