Asterio Gonzalez

Key Responsibilities:

• Team Building: Assembled and led a distributed team of 20+ members from scratch, with expertise ranging from AI/ML to AR/VR.
• Technical Leadership: Guided the design and selection of technologies for internal tools and project components.
• Proposal Development: Authored architectural and technical design sections for competitive tender proposals.
• Innovation Support: Innovation Department support to design solutions, emerging technologies evaluation, and assess their maturity.

Notable Projects

Centesimal Smart City Project (City Council of Málaga):

• Co-financed by Red.es through the Programa Operativo Plurirregional de España (POPE) and funded by the European Regional Development Fund (ERDF). Reference: 079/21-SP. https://ted.europa.eu/en/notice/-/detail/102132-2022
• Designed a comprehensive IoT infrastructure based on Kubernetes, replacing BalenaOS alternative for the Málaga Smart City initiative. The system supports over 200 buildings, featuring Wireguard VPN, secure over-the-air (OTA) updates, and offline self-operation capabilities.
• Developed a Distributed Streamed Workflow Engine for data acquisition, normalization, and KPI generation. Key features include:
  On-the-fly data normalization and resynchronization.
  Automatic generation of FIWARE/Schema.org models using Pydantic and integration with Orion Context-Broker.
  Utilized SurrealDB as the core database and broker infrastructure.

GeDia Project (CTA + RED.es):

• Designed a Digital Twin architecture for predictive models generation using Agent Simulation and Monte Carlo methods to support decision-making in Smart City infrastructure management.
• Parameterized Model Reduction techniques inspired by the research of Karen E. Willcox.
• Time series advanced prediction models, including SARIMA (Seasonal AutoRegressive Integrated Moving Average), Holt-Winters Seasonal Method, and Prophet.
• Lightweight protocol between the Digital Twin core and a 3D interface built with the Unreal Engine for VR/AR experiences. Use Unreal Pixel Streaming for web browsers remote interaction.

Smart Costablanca (City Council of Alicante):

• Co-financed by Red.es through the Programa Operativo Plurirregional de España (POPE) and funded by the European Regional Development Fund (ERDF). Reference: 010/21-SP.
• Designed and developed a planning and recommendation system for tourists, tailored to individual preferences, group composition, and typology.
  Matched user preferences with existing offers, events, and points of interest.
  Integrated Multi-Markov chains models and resource planning allocation techniques to optimize recommendations.
• Designed an indoor and outdoor tourist tracking system using beacons deployed in museums, countryside routes, and other points of interest, providing rich, location-based content experiences to enhance visitor engagement.
• Enabled real-time beacon state monitoring and configuration through the visitor's mobile application, which acted as a distributed gateway.
• Delivered a seamless and interactive experience for tourists, offering contextual information and personalized content based on their location.

Personal Researching Interests:

• Q-Learning Models: Exploring Q-Learning for software planning estimation based on digital team models.
• Logistic Regression: Investigating the use of NURBS (Non-Uniform Rational B-Splines) curves as an alternative to traditional Sigmoid-based neural network models.
  Aimed at developing a solution to better retain learned knowledge during subsequent training on unseen data, minimizing the need for full retraining.
  Focused on optimizing the training process by exposing only the relevant parts of the neuron affected by new data.
• LLM-Based Agents: Developing task automation solutions using LangChain and PydanticAI.
• Rust for AI/ML: Leveraging Rust for extensive computational tasks, edge computing, edge AI, and embedded machine learning applications.

AI/ML for Automated Trading Systems (SP500 & Nasdaq Derivatives):

• Developed advanced trading systems using AI/ML techniques, including custom preprocessing and convolutional layers for real-time data streaming and signal filtering.
• Implemented Q-Learning algorithms for autonomous trading strategies.
• Utilized Python and Dask for parallelized data processing, alongside C# for 3rd party system interoperability.
• Designed a distributed P2P computing architecture to ensure fault tolerance and resilience in trading operations across small remote devices.

Automated Learning and Imitation System for Email Classification & Routing:

• Developed an intelligent system for the company Europa, specializing in fast customs processing, to automate the classification and routing of first-level email entries.
• Analyzed and processed hundreds of daily emails, extracting and interpreting natural language data to streamline operations.
• Use of Semantic clustering techniques.
• Seamless integration with IMAP to learn from the actions of first-level email support managers, enabling automatic message routing and real-time workload balancing.
• Python / NLTK.
• Enhanced efficiency and support for timely responses in a time-critical environment.

Smart Data Extractor for IAPH (Public Organization for the Andalusian Historical Heritage Administration):

• Automated metadata extraction system from repositories of "Anuario Andaluz de Arqueología" research articles collection, spanning from 1985 to 2021.
• Approximately 3,500 full-text documents in three different languages.
• Accurate extraction and classification of locations, researcher names, dates, activities, abstracts, and summaries.
• Machine Learning and NLTK.
• Design of an alternative optimized data space for data transformation and training.
• Achieved 99.9% accuracy.

Smart Data Aggregator for IAPH:

• Development of a comprehensive system for gathering, sanitizing, and consolidating multiple data sources.
• Integration and consolidation of CSV files, digital images, and information from external websites.
• Selective web data scraping using Selenium.
• Automated selective verification and data synchronization using Selenium with the production DSpace repository to reduce inconsistencies and prevent data corruption in a live production environment.
• Python / NLTK, Selenium.

Software Factory for a Distributed, Secure P2P Collaboration Platform for "I&C Asociados"

This project involved developing a decentralized professional collaboration platform for "I&C Asociados," designed to support their distributed workforce.

Built with Python, JavaScript, and C++, the platform emphasizes security, scalability, and user-friendly collaboration tools, including Symbian OS mobile integration, ensuring seamless functionality and accessibility for users on mobile devices.

The outcome was a robust solution that enabled secure, efficient collaboration across geographically dispersed teams, meeting the holding company's needs for privacy, flexibility, and productivity.

• Decentralized P2P Architecture: ensures reliability and scalability with no single point of failure.
• Security-First Approach: end-to-end encryption, advanced authentication, and secure protocols.
• Collaboration Tools: real-time messaging, file sharing, task management, and video conferencing.
• Scalable & User-Centric: modular design for easy expansion and an intuitive interface for non-technical users adoption.
• Symbian OS Mobile Integration: extended platform accessibility specially to the project managers and sales groups.

Key Responsibilities:

• Led the development of a cutting-edge platform and SDK for next-generation slot machines, designed for Conei/Giga Games.
• Use of Java, Swing, Python, and Linux as technological stack.
• Use of a small factor PC solution that meets the hardware requirements and long term running operational stability.
• Led the development of the Fantastic League platform, a virtual football match video content generator for Sky-Bet (UK).
• Established an efficient methodology to scale the team size from 15 to 96 employees, recruiting talent from across Europe and Russia. Ensured smooth onboarding, effective resource allocation, talent retention, and sustained productivity during a period of rapid growth.
• Co-developed the "smart-stitch" optimization algorithm, which intelligently merged pre-generated animation sequences to produce 3-minute match videos.
  These videos were dynamically aligned with betting system outcomes, increasing the betting options ratio per time slot.
• Leveraged tools such as Rational Rose, Python, Trac, Bluefish video cards, and certified random hardware generators.

• Lead the development of a new competition game system for the Campus Party, the largest Spanish Internet event at the time (3000+ participants).
• Implemented QA and CI procedures.
• Replacing PHP for Python for backend processes.
• Introducing technologies and develop procedures for a well skilled but disorganized team.
• Designed and executed strategies for fast server deployment and replication, over 10+ gaming servers optimizing system performance and reliability in the Campus Party
• Utilized tools such as Rational Rose, Python, Norton Ghost, and Trac to support development, deployment, and project management processes.

• Led the Telefónica Software Factory for the development of the first version of the "112" emergency system, deployed for 112 Andalucía and 112 Galicia, along with other similar projects.
  Oversaw the design, implementation, and optimization of critical systems for emergency response.
  Designed a centralized management system for handling distributed call recording, built on top of Racal systems, to improve data accessibility for emergency services.
• Used of Rational Rose, Python, Oracle, Siemens CTI, and Racal Recorders as technological stack.
• Design of a Thermodynamic Simulation Model to optimize the performance of new air conditioning units for Hispacol.
  The model focused on ensuring efficient integration and conditioning of these units within new bus designs.
  Provide insights for a tailored driven design and point-of-failure predictions.

• Co-founded a startup focused on developing a scalable-parallel micro-chip architecture inspired by transputer concepts, in collaboration with the National Microelectronics Center (CNM) in Seville.
• Designed a full microprocessor and architecture simulator to validate the new chip design.
• Developed Rational Rose plugins in Python to enable automated code generation from UML diagrams.
• Created GCC frontend extensions to support cross-compilation of high-level languages (e.g., C++) for the new architecture.
• Managed system administration tasks using Debian Linux for the internal develop infrastructure.

Spanish Frigate F-100 / Bazán:

• Led the development of a kernel library that provided a common services layer for various modules onboard the F-100, including Electronic Warfare, Submarine Warfare, and Monitoring & Control systems.
• Ensured seamless integration with the US Navy's ATOE/AEGIS system, ensuring seamless interoperability and compliance with defense standards.
• Designed and implemented HA and Redundancy solutions using dual PowerPC nodes onboard the ship, ensuring real-time fault tolerance in the event of the master node failure.
• Utilized C++, Ada, HP-UX, and Softbench as technology stack.
• My main achieve was using real-time systems (HP-RT) to meet the demanding requirements for military-naval operations.

Alejandro Magno / Telecotrans-Fenadismer:

• Led the development of a truck fleet monitoring system utilizing the Inmarsat satellite constellation, GIS and custom hardware terminals installed onboard the trucks, enabling real-time tracking and operational insights.
• Uses C++, Informix, GIS, and Sun/Solaris, RTC, AX25 as technological stack.
• My main focus was the integration of satellite communication, geospatial data, and custom hardware to provide real-time monitoring and operational insights for terrestrial fleet operators.