Rubén De La Fuente

Developer Platform Organization — design and deployment of LLM-based systems for operational analytics, regulatory compliance, and internal tools in a regulated fintech environment.

AGENTIC SYSTEMS

I designed and deployed an agentic analytics platform in production using the Anthropic API (Claude tool use), enabling natural language queries over BigQuery — full ownership of the lifecycle, from problem definition to stakeholder adoption. I implemented prompt versioning, iterative optimization, and evaluation harnesses to continuously measure output quality and reliability.

LLM EVALUATION AND STRUCTURED EXTRACTION

I developed a batch LLM extraction pipeline that transforms unstructured P1+ incident reports into structured compliance datasets — adopted by the risk team for operational analysis. I designed schema-based JSON extraction with Claude tool use, building QA and validation workflows with pytest to ensure auditability of over 100 critical documents. I developed evaluation frameworks based on my Machine Translation experience (BLEU, TER, formal metrics) adapted to assess LLM output fidelity, consistency, and regulatory alignment.

ENTERPRISE AI TOOLS

I developed reusable LLM-powered tools (documentation automation, EDA assistant, survey design) integrated into the corporate repository — contributing to AI governance and best practice standards. I collaborated with Directors, VPs, and Principal Engineers to translate AI capabilities into operational solutions; I mentored junior data scientists and engineers.

Over 7 years of hands-on experience in multilingual NLP and language technology, directly applicable to enterprise language solutions.

I led the end-to-end expansion of PayPal's Machine Translation infrastructure to over 25 languages — encompassing translation pipeline design, model integration, linguistic QA, and production deployment. I evaluated multilingual model quality using formal MT metrics (BLEU, TER, quality estimation), performing systematic error analyses and iterating on model performance. I designed automated NLP preprocessing pipelines and built datasets for testing and QA of new model versions — including annotation guidelines and quality control processes. I integrated translation APIs into internal systems; served as an internal language technology consultant and Subject Matter Expert (SME) for product and engineering teams.