Seville, Andalusia, Spain
Hi! I’m Antonio Campuzano. My journey as a software engineer started in 2018 when I entered the Software Engineering degree. Then my professional career started at GESI S.L. in 2020, while I was only in my second year of university. This allowed me to connect academic theory with real-world problems from the very beginning. Since then, I’ve had the chance to work in international environments: building systems at ArcelorMittal (France), leading backend development at Integrated Worlds (Germany), and now working as a Software Engineer at Rithum (United States). Throughout these years, I’ve focused on making systems scalable, reliable, and efficient. After years of designing microservices and cloud infrastructure, I’m now moving towards Applied Generative AI. My goal isn't just to "chat" with models, but to use my engineering background to build the infrastructure that makes AI useful for businesses. What I’m working on today: AI Engineering: Designing RAG (Retrieval-Augmented Generation) systems and using Vector Databases to give models the right context. Infrastructure: Deploying and managing LLMs , making sure they are secure and ready for production. Solid Foundations: I still rely on my experience with .NET, C#, and Microservices, because I believe great AI needs a strong backend to work properly. I’m looking to bridge the gap between traditional software engineering and the new world of LLMs. If you’re looking for an engineer who understands both backend architecture and AI implementation, let’s talk!
I operate at the intersection of robust Backend Engineering and Applied AI. My core focus is ensuring reliability for high-volume e-commerce operations. To support this scale, I designed and operate a production machine-learning forecasting platform for ad-spend-to-revenue transformation. By routing over 460 client profiles and tracking $2.3B in revenue, I ensure we maintain strict accuracy thresholds using a robust model cascade that includes Temporal Fusion Transformers, LightGBM, and AutoARIMA. I engineered our end-to-end MLOps pipeline natively on AWS using SageMaker, EKS, and S3. Identifying a gap in user experience, I also proposed and currently lead the development of a multi-agent GenAI assistant. I architected this production-grade pipeline—featuring a Planner, Router, Evaluator, and Executor—to allow non-technical users to generate complex business logic using natural language. Unlike typical demos, I engineered this solution using advanced Retrieval-Augmented Generation (RAG) and Qdrant for hybrid vector search, leveraging both self-hosted (vLLM) and Azure LLMs to ground the models in dynamic business context. To ensure professional standards, I treat AI implementation with the same rigor as our traditional software: I built a custom automated evaluation framework and real-time Grafana dashboards to strictly monitor the model's accuracy, latency, and token usage in production
As the primary backend engineer and cloud architect, I owned the full lifecycle of our applications, from initial architecture to managing multi-environment CI/CD deployments in Azure. I solely engineered the backend of a multi-tenant, white-label enterprise survey platform that successfully serves over 1,000 users. To scale this securely, I implemented enterprise SSO via Microsoft Entra ID, completely eliminating manual per-user onboarding while ensuring strict data isolation across clients. A major focus of my role was modernizing our infrastructure. I rebuilt a legacy EDI integration platform from the ground up as a robust system of containerized .NET microservices. By leveraging RabbitMQ topic exchanges for routing and failure isolation, this new event-driven architecture seamlessly processes over 100,000 documents daily, transforming data between clients with mismatched schemas. I also actively looked for ways to bridge traditional backend logic with modern AI. I engineered a custom email-response handler that paired deterministic SMTP bounce parsing with an AI classifier, interpreting complex free-text replies to flag unreachable contacts and successfully prevent domain blacklisting.
As a software engineer, I participated in the backend development of a large-scale, multi-team industrial platform. My work specifically centered on the platform's core quality-assurance module. Working within a modular microservices architecture, I developed and maintained the complex backend logic required to officially certify that industrial materials met strict, predefined quality specifications. I drove feature development from end to end, deploying these highly reliable, containerized solutions using C#/.NET, Docker, and Azure. Additionally, I managed our CI/CD pipelines in Azure DevOps to optimize code delivery across multiple environments, ensuring seamless integration and high system performance for the end users.
During my tenure at GESI, I honed my skills in software development, primarily focusing on Object-Oriented Programming (OOP), SOAP, and various web services including HTTP. I actively contributed to projects utilizing Scrum methodologies and Agile frameworks, ensuring efficient and timely delivery of software solutions. Key Responsibilities and Achievements: Developed and maintained robust software solutions using Microsoft SQL Server, Spring Boot, MySQL, HTML, Spring Framework, CSS, and Python. Proficiently used Entity Framework, Microsoft Azure, .NET Framework, SQL, Git, C#, and AngularJS for various development tasks. Engaged in full-stack software development, demonstrating strong problem-solving skills and programming acumen. Collaborated effectively in a team environment, contributing to the continuous improvement of development processes and practices. This role at GESI was instrumental in refining my technical expertise and professional growth, laying a strong foundation for my career in software development.