Process immersion
We map the real flow, the exceptions, the human actors, and the decision points before writing a single line of code.
We build machine learning and LLM systems to automate corporate processes of the highest complexity — from the front-end to proprietary models.
In production with
The most relevant players in the sectors that transform Brazil.
We work where critical processes demand more than a plug-in model: they involve regulation, contracts, unstructured documents, human exceptions, and decisions with financial impact.
We dive into the partner company's target process until we understand it as well as those who run it. From there we design the entire application — from the front-end to OCR microservices, embedding models, LLM agents, and proprietary review mechanisms.
We map the real flow, the exceptions, the human actors, and the decision points before writing a single line of code.
Senior teams shipping production products — not disposable MVPs. Interfaces, distributed back-ends, OCR pipelines, RAG, embeddings, agents, and proprietary models.
The IP of what we build is ours. Proven with a reference partner, then scaled across the entire sector.
We work with sector-leading companies to automate critical processes that have already proven hard to solve in-house or through traditional vendors. The partner company brings the problem, the operational context, and the domain knowledge. We build, operate, and evolve the technology — with a focus on product, scale, and real impact.
We don't sell hours or build-on-demand squads. We come in as a long-term tech partner, until we understand the process — its rules, exceptions, documents, and decisions. The result: a mature, production-ready solution, in less time, with less risk, and for a fraction of the cost of other vendors.
We pursue specific, recurring problems where simple automation isn't enough. Each deployment sharpens the product, the models, and the sector knowledge — and the gain is mutual: every partner benefits from what we learn with the others, including the first one. The solution gets more robust with every new client, for everyone.
Each product is born inside a specific, market-leading partner and matures as replicable technology for the whole sector.
Sector
Higher Education
Automation of academic transfers. Advanced OCR for transcript extraction, NER and embeddings to structure courses, and NLP for exemption suggestions, with proprietary LLM agents reviewing the process.
Partners
Cogna
Ânima
YDUQS
Sector
Higher Education
Platform for end-to-end automation of academic flexibility in higher education. Machine-learning recommender models help each student find the best courses for their academic moment, and linear-optimization algorithms maximize the operational efficiency of students per classroom.
Partners
ESPM
Higher education institutions
Sector
Hospital Care
Automated defense of hospital claim denials. Searches contractual bases for legal-technical grounds, building appeals with RAG, embedding models, and specialized agents.
Partners
Hospital Sírio-Libanês
Sector
Labor Law
Processes thousands of labor lawsuits extracting relevant information and entities (NER), embeddings for RAG, and regression models that estimate the likely return value of each case.
Partners
Corporate legal departments
Sector
Publishers
Automatic book ingestion with metadata extraction — chapters, table of contents, entities. Lets readers chat with the book through any messaging tool, including WhatsApp, using RAG and proprietary agents.
Partners
Editora do Brasil
We build and operate each application end-to-end — from front-end to models, through architecture and DevOps. Clients consume it as SaaS.
LLM Models
Proprietary agents
ML Models
Clustering, regression, classification
NLP
NER & classification
Computer Vision
Detection & image classification
Retrieval
RAG & Embeddings
Documents
Advanced OCR
Recommendation
ML-based systems
Optimization
Linear programming & constraints
Architecture
Monolith + microservices
Front-end
Web apps or chat interfaces
BI
Dashboards & visualizations
SaaS Operations
DevOps, scale & observability
1M+
High-complexity items processed in our pipelines.
4
Distinct sectors running our technology at scale.
8 years
Operating in the market.
That's exactly where we want to be.
Tell us about the problem — if there's fit, we'll set up a long-term partnership.
What we evaluate