Christopher Manzano Vimos

Iberdrola — renewable-energy forecasting (AWS)

• Designed distributed ML pipelines forecasting output for 200+ renewable plants across the US, Spain and the UK, ingesting 2 TB of data daily.
• Built ingestion and normalization of meteorological data from ECMWF, GFS and WRF, engineered for reproducibility and robustness in production.

Intermundial — LLM claims automation

• Architected LangChain/LangGraph pipelines for hierarchical document classification and information extraction, reaching over 90% accuracy on identity fields (DNI, NIE, passport).
• Built an LLM-as-judge evaluation framework and deployed open-source LLMs on private infrastructure under data-privacy constraints.

Nexudus — predictive models & GenAI (SaaS)

• Churn-prediction and occupancy-forecasting models driving retention, pricing and capacity decisions.
• Led the Generative-AI workflows behind automated content for the Flexspace Observatory blog.

Bid Genius — LLM assistant for public tendering

• Early-stage build of a proprietary LLM tool for assisted generation of tender documentation — solving long-document coherence and context management.

JOT Internet Media — NLP keyword classification

• Multi-stage hierarchical classifier based on RoBERTa for Google Ads keyword ranking, fine-tuning Hugging Face models on digital-marketing data.

Fen (US Department of Defense) — medical AI research

• Computer-vision models for hernia detection, vertebral segmentation and scoliosis/lordosis detection to support radiologists, delivered as a clinical web application.