Johsac Isbac Gomez Sanchez
PhD Researcher in GenAI & Large Language Models | AI Agent | Statistician & Data Scientist | Python, PyTorch, RAG, LangChain
Campinas, Brazil johsacgomez16@gmail.com
About
PhD Researcher in Generative AI & Large Language Models at the Laboratory of Computing and Automation (LCA), Universidade Estadual de Campinas (UNICAMP), Brazil. My research focuses on multimodal generative architectures for 3D gesture synthesis driven by speech and text — winner of the Reproducibility Award at GENEA 2025 (ACM Multimedia).
With a Master's in Statistics (UNICAMP, 2023) and a background in Mechanical Engineering (ESPOL, 2020), I bridge the gap between statistical rigor and AI innovation. My stack spans RAG, LangChain, PyTorch, Agentic AI, and I'm passionate about building production-ready AI systems that serve both academic research and industry applications in Tech, FinTech, and Consulting.
Awards & Honors
Reproducibility Award — GENEA 2025
ACM Multimedia (ACMMM 2025) • GENEA Workshop on Gesture Generation
Awarded for the LLM-Gesture-Pipeline work demonstrating reproducible, state-of-the-art gesture generation from speech using Large Language Models.
Experience
PhD Researcher — Laboratory of Computing and Automation (LCA)
Investigador de Doctorado — Laboratorio de Computación y Automatización (LCA)
Mar 2023 — PresentUNICAMP, Campinas, Brazil
Researching multimodal generative architectures for 3D virtual character gesture synthesis. Developing LLM-driven pipelines integrating speech, text, and motion data. Winner of the Reproducibility Award at GENEA 2025 (ACMMM).
Academic Researcher & Teaching Assistant
Investigador Académico & Profesor Asistente
2023 — 2024UNICAMP, Campinas, Brazil
Applied research in Machine Learning, Generative AI, and Statistics. Teaching assistant for courses in statistics and computational methods.
Master's Researcher — Institute of Statistics
Investigador de Maestría — Instituto de Estadística
Mar 2021 — Oct 2023UNICAMP, Campinas, Brazil
Developed Markovian models, metrics in state spaces, and simulations with Markov chains. Published a novel metric in the journal Entropy.
Education
PhD in Electrical Engineering (Computer Science — GenAI & LLMs)
Doctorado en Ingeniería Eléctrica (Computación — IA Generativa & LLMs)
Mar 2023 — PresentUNICAMP, Campinas, Brazil
Research at the Laboratory of Computing and Automation (LCA). Focus on NLP, Generative AI, LLMs, and multimodal architectures.
Master in Statistics
Maestría en Estadística
Mar 2021 — Oct 2023UNICAMP, Campinas, Brazil
Specialization in computational statistical methods, data analysis, Markovian processes, and machine learning algorithms.
Bachelor in Mechanical Engineering
Ingeniería Mecánica
Sep 2014 — Sep 2020ESPOL, Guayaquil, Ecuador
Thesis: Analysis and design of an improved green coffee transportation system.
Publications
From Embeddings to Language Models: A Comparative Analysis of Feature Extractors for Text-Only and Multimodal Gesture Generation
Benchmarking Speech-Driven Gesture Generation Models for Generalization to Unseen Voices and Noisy Environments
Featured Projects
LLM-Gesture-Pipelines
LLM-driven pipeline for generating realistic 3D body gestures from speech. GENEA 2025 entry winning the Reproducibility Award at ACM Multimedia.
Benchmarking-SDGG-Models
Framework for evaluating Speech-Driven Gesture Generation models across unseen voices and noisy environments.
Systematic Review Automator
AI-powered tool that automates metadata extraction from scientific papers into Google Sheets using Gemini API and agentic AI workflows.
eToro Portfolio Evolution
Professional-grade financial analytics dashboard with AI Copilot powered by Gemini. Multi-agent orchestrator for portfolio analysis.