Data Science 
and 
Machine Learning 

Hello, I'm Luis Sante. Data Scientist Researcher with advanced Software Engineering expertise, building scalable AI products from research to production.

Natural Language ProcessingGraph Neural NetworkTransformersGraphRAGFullStack
Luis Sante

Experience

Mar 2025 - now

LegalXplain

Backend Developer

  • Developed backend services for LegalXplain platform using NLP
  • Deployed scalable ML/NLP pipelines into production
  • Improved accessibility and understanding of legal documents
Visual Data Science - FGV EMAP

Rio de Janeiro, Brazil

Oct 2024 - Feb 2025

Sentiment Analysis and Structural Topic Modeling (SA-STM)

Machine Learning Researcher

  • Applied sentiment analysis and Structural Topic Modeling (STM)
  • Classified software requirements from user reviews
Excelsius

Lima, Peru

Jun 2024 - Oct 2024

Trading World EVS Official Website

Freelance Full-Stack Developer

  • Developed official website
  • Built responsive UI and backend integration
Freelance

Arequipa, Peru

Oct 2023 - Oct 2024

LexCom

Tech Lead

  • Led development of profitability prediction system
  • Built ML models for e-commerce data
  • Integrated OpenAI-based text generation services
  • Handled full-stack and cloud deployment
Lexcom.tech

Arequipa, Peru

Publications

UrbanClipAtlas: A Visual Analytics Framework for Event and Scene Retrieval in Urban Videos

Data Scientist Researcher - Visual Data Science - FGV EMAP

Conference: EuroVis

Knowledge

Vision-Language Models (VLMs)Knowledge GraphsSpatiotemporal Data AnalysisEvent and Scene RetrievalComputer Vision

Technology

PythonHuggingFaceYOLOv8-WorldSvelteTypeScriptWebSockets

STRive: An association rule-based system for the exploration of spatiotemporal categorical data

Data Scientist Researcher - Visual Data Science - FGV EMAP

Journal: Computer & Graphics - SIBGRAPI

Knowledge

Association Rule MiningClusteringSpatiotemporal Data Analysis

Technology

Pythonscikit-learnD3.jsTypeScript

Projects

Security Urban Perception Tracking

Visual prototype exploring urban security perception using eye-tracking data across participants.

Date: Jun 2025

Location: Rio de Janeiro, Brazil

Knowledge

Eye-Tracking Data AnalysisVisual AnalyticsSpatiotemporal Data Visualization
Fine-Tuning LLaMA 2 with QLoRA

Conducted fine-tuning of the 'NousResearch/Llama-2-7b-chat-hf' model using the 'mlabonne/guanaco-llama2-1k' dataset to create the customized 'llama-2-7b-miniguanaco' model. Implemented QLoRA with a 64-dimension LoRA attention layer and 4-bit precision to optimize model performance while maintaining resource efficiency.

Date: Aug 2024

Location: Arequipa, Peru

Knowledge

Fine TuningPythonHuggingFacePyTorchLLaMA2QLoRA

Bachelor's Thesis

Topic modeling using Agglomerative Clustering and Sentence-BERT on user comments

Topic modeling to find functionalities within user reviews of a Google Play Store app using unsupervised algorithms and sentence embeddings with Transformers. This research contributed to better understanding user needs and app development priorities.

Location: Arequipa, Peru

Date: Aug 2024

Knowledge

NLPTopic ModellingTransformersClustering

Contact me