Projects
The following are a series of projects developed by the chair.
Introduction
This section will include everything that research project that has been transferred to the business sector and is currently used to solve problems or improve processes and decision-making.
Among the objectives of the chair is Promote the connection between the university community and the business sector Through research projects that generate scientific and technological knowledge. By disseminating and transferring this knowledge to society and the productive sector, the aim is to promote and strengthen business activities related to Big Data, Open Data, and Blockchain.
The benefits of applying Big Data, Open Data and Blockchain in companies include, among others, improvements in: process effectiveness and decision-making, efficiency and cost optimization, customer segmentation, data security, accessibility to information within the company, new sources of income and competitive advantages.
SPC-19
This project was carried out in collaboration with Mutua Tinerfeña, who proposed the challenge of offer insurance that covers the possibility of a tourist contracting COVID-19.
Since the parameters used for this type of insurance must be objective, transparent, and consistent, the use of databases and their analysis become crucial in processing these policies, and that's where blockchain technology comes in. This technology allows for the secure exchange of data without the need for intermediaries, making it especially useful for all types of online transactions, not just financial ones.
Local tourism has been severely impacted by the pandemic, and incentive measures are needed to gradually recover this important engine of the Canary Islands' economy. The SPC19 project emerged precisely as a proof of concept to restore tourist confidence when visiting the Canary Islands. Through well-defined processes, the The insurance company, the hotel, and the PCR testing laboratory are collaborating to offer tourists compensation in the event of contracting the disease.. The use of smart contracts is effective for implementing parametric insurance, as it facilitates the automatic management of compensation in case of contagion in the islands.
Blockchain management makes it easier for each node in the system (insurance company, hotels, and laboratories) to have only the necessary information available for processing, which ensures data protection for each party and also speeds up the process.
Implementation of a text anonymization and classification system based on large language models in the Tenerife Island Council.
This project has been carried out in collaboration with the Tenerife Island Council, institution that posed the challenge of modernizing and optimizing document management and citizen services at its electronic headquarters.
Given that public administrations handle on a daily basis large volumes of information, The secure and efficient processing of data is of great importance, and that's where machine learning techniques and Large Language Models (LLMs). This technology allows for the agile processing, understanding, and categorization of complex texts, making it especially useful for automating workflows that traditionally require significant manual effort.
The volume of citizen requests necessitates innovative tools to expedite response times and improve public service. This project arises precisely as a technical solution to provide the administration with a greater efficiency. Through the training an Artificial Intelligence model with the historical records, The system is able to intelligently classify requests, automatically identifying the corresponding processing flows according to the topic raised by the citizen.
The implementation of this system stands out for incorporating a text anonymization service that operates in a strictly local. By operating without the need to send information to cloud services, the technological infrastructure ensures that personal data does not leave the institution, guaranteeing data sovereignty and adherence to data processing principles. This protects the privacy of each party involved and, furthermore, securely streamlines the processes of the electronic headquarters.
Project dates:
21/12/2023 – 21/06/2024
Knowledge transfer for the development of solutions based on natural language processing and process automation.
This project was carried out in collaboration with the company MIA SL., an entity that posed the challenge of optimizing the complex management of contracts and billing cycles for services for large telecommunications operators.
Given that these companies manage a massive volume of documentation In commercial transactions, information extraction and its analytical cross-referencing become very important, and that's where the techniques of Natural Language Processing (PLN). This technology allows digitization and understand contractual texts complex in a structured way, making it especially useful for automating validations that traditionally require arduous, error-prone manual processes.
The large number of clients and the diversity of contracted services necessitate innovative tools to ensure accurate billing and avoid discrepancies. This project arises precisely as a technical solution to equip the operators of a elderly operational reliability. Through advanced data processing algorithms, The system is able to cross-reference the information extracted from digitized contracts with service billing data.
The implementation of this system facilitates that can identify systematically the differences among the agreed and it charged. This allows for the early detection of errors and the automated validation of services provided, which ensures greater accuracy in the revenue cycle for the company and also drastically streamlines internal audit processes in telecommunications.
Project dates:
7/3/2023 – 31/07/2023
Application of automation and robotization techniques to processes in entities of the financial sector: analysis, design, implementation of use cases and measurement of results obtained after their production.
This project has been carried out in collaboration with Cajasiete, an entity that posed the challenge of transforming its daily operations through the automation of key organizational processes.
Given that financial institutions require a high degree of efficiency in their operations, the standardization and optimization of workflows become very important, and that's where the Robotic Process Automation (RPA). This technology allows repetitive tasks to be executed in a structured and unattended manner, making it especially useful for streamlining banking services and drastically reducing manual workload.
Adopting these new tools requires significant organizational changes and a methodical approach. This project arises precisely as a solution comprehensive for guide this technological and cultural transition in a structured way. Through the analysis and identification of the entity's key processes, the necessary technical milestones were achieved for the deployment and production of the most suitable robotics tools.
The implementation of this system makes it easier for the institution not only to automate its tasks, but also consolidate a new culture organizational well-founded in the take of decisions based in data. This allows for the transparent quantification of benefits obtained and, furthermore, through the use of indicators and dashboards, establishes a methodology for constant measurement to ensure the continuous improvement of operations.
Project dates:
23/12/2022 – 22/12/2023
New actions within the Sustainable Mobility Strategy initiative in the Anaga Biosphere Reserve. Data and Smart Mobility Strategy 2025–2026
This project has been carried out in collaboration with the Tenerife Island Council, institution that posed the challenge of strengthening territorial planning and optimizing strategic decision-making in matters of mobility.
Since transport management and route analysis require the integration of multiple information sources, data protection and secure processing are of paramount importance, and this is where anonymization techniques and Artificial Intelligence come in. This technological architecture allows Enrich mobility data with external sources strictly complying with the GDPR, therefore it is especially useful for processing glarge volumes of vehicle information without compromising the privacy of the citizenship.
The constant need to organize the movement of residents and visitors necessitates innovative tools to facilitate the use of public transport and anticipate traffic problems. This project arises precisely as a technical solution for trigger key assets in the land management. Through the use of simulators and analytical processes, along with the development of a integrated smart assistant with GTFS data, The system is capable of rto answer queries about routes and schedules using natural language, as well as establishing methodologies to identify circulation patterns.
The implementation of this ecosystem makes it easier for the administration to anticipate risks of saturation on the roads based on impact assessments and rigorous feasibility studies. This ensures secure integration of information and, moreover, dramatically speeds up access to mobility and public transport data for end users.
Project dates:
15/10/2025 – 20/6/2026
Implementation of RAG agents for university management and entrepreneurship
This project has been carried out in collaboration with the University of La Laguna (ULL) and his the General Foundation of the University of La Laguna (FGULL), institutions that posed the challenge of helping technicians and modernizing the attention to future students and entrepreneurs through automated solutions.
Given that academic admissions processes and the development of new businesses generate a high volume of complex questions, the delivery of accurate and reliable information becomes extremely important, and that's where the conversational assistants advanced based on RAG architecture (Recovery-Augmented Generation). This technology allows processing natural language queries restricting responses exclusively to validated sources, Therefore, it is especially useful for preventing any kind of deviation and providing certainty in institutional environments.
The constant need to guide users through their procedures necessitates innovative tools to reduce waiting times and optimize resources. This project arises precisely as a technical solution to provide both entities with channels through which to...Specialized technicians can reduce your workload. Through the deployment of the «ULL Access Agent» and the «FGULL Entrepreneurship Agent«"The system is capable of resolving issues quickly." queries about the university tuition and provide Specialized advice to promote new business projects.
The implementation of this ecosystem allows us to offer a Total accuracy guarantee, ensuring that all answers are generated strictly from the official documentation injected into the system and, in addition, it significantly streamlines the workload of the university and foundation's customer service teams.
Project dates:
14/6/2025 – XX
Automatic Occupancy Coding (CNO) using LLMs
This project has been carried out in collaboration with the Canary Islands Institute of Statistics (ISTAC), the entity that posed the challenge of Automate the coding of open-ended responses in citizen surveys towards the National Occupational Code (CNO).
Since public surveys collect very diverse responses and require standardized and accurate categorization, advanced semantic analysis becomes very important, and that's where the techniques of Machine Learning and the Large Language Models (LLMs). This technology allows for understanding the context and nuances of job descriptions provided by citizens, making it especially useful for process of shape intelligent large volumes of text that traditionally require arduous manual sorting.
The high volume of statistical data processed periodically necessitates innovative tools to optimize public resources and reduce processing times. This project arises precisely as a technical solution to equip the epublic statistics of greater efficiency. Through the integration of key metadata (such as education level or employment status) and the use of hybrid approaches with embeddings and efficient fine-tuning (PEFT), the system is able to capture semantic nuances that conventional classification methods usually ignore.
The implementation of this infrastructure allows the organization to process its surveys with a high degree of autonomy. This enables the automation of Safely form the 86% of the sorting volume with a margin of error of 10%, This consolidates a robust technological solution and also drastically speeds up the generation of reliable public statistical data.
Project dates:
31/10/2023 – 30/10/2025
On-Premise Document Anonymization System
This project has been carried out in collaboration with Cajasiete, an entity that posed the challenge of securely processing highly sensitive information while complying with the strictest requirements regulatory and data sovereignty, like the GDPR and the LOPDGDD.
Since the handling of documents containing Personally Identifiable Information (PII) and Protected Health Information (PHI) prohibits sending plain text to external providers, the local infrastructure charges one great importance, And that's where on-premise processing strategies and language models (LLMs) come in. This technology allows for the autonomous detection, replacement, and physical destruction of sensitive data in various formats (PDF, Word, and plain text), making it especially useful for to guarantee privacy without compromising the institution's operational capacity.
The risk of regulatory non-compliance necessitates innovative tools that maximize the thoroughness in the detection, assuming that over-detection is preferable to a false negative. This project arises precisely as a technical solution to provide the entity of a shielded work environment. Through a system that offers two interchangeable strategies, a multi-phase hybrid pipeline, and an LLM-driven multimodal approach, the system is able to accurately anonymize standard text documents.
Implementing this architecture makes it easier for the organization to obtain, for each document processed, a clean, metadata-free version along with a detailed audit report. This allows for technical traceability of all detected entities, their coordinates, and the generated substitute, ensuring regulatory rigor and drastically streamlining data protection assessment and compliance processes.
Project dates:
25/5/2026 – XX
