AI Engineer & Researcher

Technical skills:
Large Language Models, Machine Learning, MLOps, Agents-LLM, RAG, Knowledge-Graphs, DocumentAI (Python, C/C++, R)

Research interests:
1. Entanglement, complexity and sensitivity in quantum sensors, NMR interface quantum computer.
2. Categorical logic, Topos Theory, Complexity Theory.

Education

- MSc Complex Systems Modeling | University of Sao Paulo (_December 2023_) Research: Quantum learning from quantum sensing experiments.

Work Experience

Data Scientist GenAI @ KT Group (Jan 2024 - Present)

LLMs development, training, fine-tuning and deployment of LLM for Document AI, OCR, NLP, Multimodality, graphs and RAG. Stack: BERT, Sentence-BERT, LayoutLM, Claude 3, OpenAI’s GPT-4, Meta’s LLaMA 2, Mistral-7B.

Data Scientist @ Learn to Fly (Jul 2023 - Dez 2023)

Performance, scaling, statistical modeling, behavioral data-science, data-driven insight and in-depth analysis for C-Level decision.
Machine Learning (gradient boosting trees/LGBM); Experiment design (A/B Test, ANOVA); Outlier/anomaly detection (isolation forest); Unsupervised learning (pagerank, clustering).

Data Scientist Consultant @ Cogna (Dec 2022 - Dez 2023)

Data Science, AI, Lectures Neural Networks, Deep Learning degree academics courses. Cogna has more than 2 million students, with 290,000 students.

Data Engineer @ Blueshift (Oct 2021 - Jan 2023)

Data Factory, provisioning resource Azure, data ingestion and integration, ELT, Datalake, processing in Databricks with Spark. Power Apps and integration with Power BI, tabular modeling, dataviz and reports, DAX.

Assistant Researcher @ IBM (Jan 2018 - Jan 2020)

Cognitive Computing, NLP , Data Science and Predictive Data Mining. Developer a biological neural network to simulate the process of recognizing verbs in the brain. Deep Learning and Data Science to classify experimental neuromage data from linguistic tasks such as semantic and verb interpretation.

Projects

LLM & NLP

Prophet: modelo forecasting em sensores de IoT

Fine tuning de modelo de embedding com Sentence-BERT RAG simples e rápido com Langchain e Pinecone no Databricks Prophet: modelo forecasting em sensores de IoT Tupy-GPT: LLM quantizado e adaptado para PT-BR Fine-tune a generative AI model for dialogue summarization Fine-tune FLAN-T5 with reinforcement learning to generate more-positive summaries DocMLLM full-cycle com Claude 3

Quantum computing

Group-theoretic and Representation theory

Computational Neuroscience

Bike Study

Talks & Lectures

Causality: The new science of an old question - GSP Seminar, Fall 2021

Publications

• Sensoriamento Quˆantico com Aplica¸c˜oes em Defesa, XXV Simp´osio de Aplica¸ c˜oes Operacionais em ´ Areas de Defes, Samuel da Silva, Alexandre Ferreira Ramos, Warley M. Santos Alves, Journal and Event Management System (JEMS), 2023. Link.

• Mirror Neurons Modeled Through Spike-Timing-Dependent Plasticity are Affected by Channelopathies Associ ated with Autism Spectrum Disorder Gabriela Antunes, Samuel F Faria da Silva, Fabio M Simoes de Souza International Journal of Neural SystemsVol. 28, No. 05, 1750058 2018. Link.

• Computational Simulation of Mirror Neurons Through of the STDP VI Scientific Initiation Meeting 2017, Federal University of ABC, Santo Andre, Sao Paulo.

• Mirror Neurons on Computational Condutance-Based Model and Implications with Autism XV Brazilian Society for Neuroscience and Behavior, 2017 Araraquara, Sao Paulo.

• Neural Mirroring in Integrate-and-Fire Model II Conference on Bioathematical, Unicamp, Campinas, Sao Paulo