Sofía Sonia Martinez

Architected and maintained distributed backend services for high-throughput automated AI evaluation pipelines. Built asynchronous, low-latency REST APIs and Redis-backed caching/queues, and designed structured LLM evaluation frameworks.

Built reactive internal dashboards for visualization of LLM training outputs and human preference data collection. Developed scalable database schemas and workflows, optimized SQL performance, and improved dataset processing performance via asynchronous background workers.

Engineered high-fidelity datasets and simulation models for training multimodal AI. Built GIS mapping and scientific visualization pipelines using QGIS and ParaView, reconstructed 3D medical datasets from DICOM, and created parametric CAD prototypes with FreeCAD.

Developed high-throughput data processing and orchestration pipelines for large ML training datasets. Implemented parallel processing for dataset parsing/validation and created internal automation tooling to test, deploy, and monitor backend evaluation workloads.

Scaled machine learning model serving microservices on Kubernetes to ensure low response latency. Implemented CI/CD automation for NLP services and cleaned/transformed enterprise databases to provide refined training assets.

PhD research focused on scalable AI systems, distributed systems architecture, and optimization of machine learning pipelines. Dissertation work centered on designing and optimizing large-scale AI-driven systems.

MSc program specializing in machine learning, cloud computing, and scalable system design. Coursework centered on building and optimizing software systems for ML workloads.

BSc degree focused on data structures & algorithms, object-oriented design, and full-stack development. Built foundational software engineering skills across core CS and application development.