Open to opportunities

Yosmely Bermúdez

@yosmelybermdez

Message

AI Engineer specializing in LLM applications and intelligent automation, from prototype to production.

Argentina

Message

What I'm looking for

I’m seeking remote roles to deploy AI models from prototype to production—building LLM-based products, agentic workflows, and document/NLP pipelines that deliver measurable impact.

I specialize in LLM applications and intelligent automation, building production-ready chatbots with RAG, NLP pipelines for document extraction, and full-stack systems. I bring a deep learning background (autoencoders, SOMs) applied to geophysical and biomedical data, and I enjoy turning messy, unstructured inputs into reliable outputs for real teams.

Most recently, I developed an LLM-powered chatbot with LangChain (Gemini, Groq, OpenAI) for strategic decision support using a company knowledge base, and it’s currently in production for internal use. I also built an automated PDF menu extraction system that reduced demand planning time by 65%, and I architected an end-to-end agentic workflow capturing text, voice messages, and PDFs—then logging structured sales data into a Notion CRM with zero manual input.

Experience

Work history, roles, and key accomplishments

AI/LLM Automation Developer

Caracas Catering

Oct 2025 - Jun 2026 (8 months)

Developed an LLM-powered chatbot with LangChain using multiple model providers and the company knowledge base for strategic decision support, deployed for internal use. Built an automated PDF menu extraction workflow that reduced demand planning time by 65%, and an agentic n8n pipeline that captured unstructured interactions and logged structured sales data into a Notion CRM with zero manual input

LLMs (LangChain RAG PDF Extraction (Pypdf Pdfplumber)REST APIs Webhooks

Machine Learning Research Engineer

FUNVISIS

Jan 2025 - Oct 2025 (9 months)

Developed an end-to-end ML pipeline for seismic trace classification across 500+ waveforms using NLP embeddings and Self-Organizing Maps (SOMs) for unsupervised clustering. Implemented automated categorization and delivered a production-ready pipeline using TensorFlow Embedding Projector for visualization and ObsPy for seismological data processing.

Python TensorFlow ObsPy Self Organizing Maps (SOMs)Embeddings TensorFlow Embedding Projector Unsupervised Learning Seismic Data Processing