Matheus Ferraroni Sanches

CTO at Orion Sistemas Agrícolas and a PhD candidate in Computer Science at UNICAMP. His research and development work spans agtech platforms, remote sensing, natural language processing, federated learning, LLM privacy, optimization, and vehicular networks.

Natural Language Processing Agtech LLM Privacy and Security Federated Learning Optimization Remote Sensing Vehicular Networks Wireless Communication

Professional experience

Chief Technology Officer (CTO), Orion Sistemas Agrícolas

Jun 2025–present

Leads Technology, Innovation, and Research & Development, contributing to the creation of new products, agricultural machinery, and digital platforms. He is responsible for the company’s technology strategy across software development, enterprise systems, infrastructure, information security, and digital transformation.

Professor, Universidade de Marília (UNIMAR)

Jan 2025–present

University professor in Artificial Intelligence, Computer Science, and Information Systems. Teaches courses related to artificial intelligence, machine learning, algorithms, and data structures, connecting theoretical foundations with practical applications to develop innovative technology solutions.

Chief Technology Officer (CTO), CYGNI AgroScience

Sep 2020–Jun 2025

Led technology strategy, software architecture, and development teams, guiding the evolution of remote sensing and geospatial analysis platforms based on satellite imagery, from product conception through operation and customer delivery.

Selected previous experience

  • SumUp · Senior Software Engineer · Jun 2022–Jun 2025
  • CI&T Software · Data Scientist/Researcher · Feb 2021–Sep 2022
  • CYGNI AgroScience · Tech Lead · Dec 2019–Aug 2020
  • Clickideia Tecnologia Educacional · Data Scientist/Researcher · Dec 2018–Nov 2019
  • CYGNI AgroScience · IT Manager · Aug 2016–Nov 2018
  • Centro Universitário Eurípides de Marília, UNIVEM · CNPq Researcher · Jan 2015–Jul 2016

Education

PhD in Computer Science

2020–present

Universidade Estadual de Campinas, UNICAMP. Advisor: Leandro Aparecido Villas.

Master’s in Computer Science

2018–2020

Universidade Estadual de Campinas, UNICAMP. Dissertation: “Alocação de Road Side Unit Ciente de Obstáculos com Diferentes Modelos de Propagação de Sinal”. Advisor: Leandro Aparecido Villas.

Bachelor’s in Computer Science

2014–2017

Centro Universitário Eurípides de Marília, UNIVEM. Undergraduate project: “Processamento e Entendimento de Linguagem Natural no Gerenciamento de Emergências Para Obtenção de Consciência Situacional”. Advisor: Leonardo Castro Botega.

Skills

  • Programming · Python, PHP, JavaScript, Elixir, C++, Go
  • Tools · Git, Docker, Google Maps JS, Jupyter, Scikit-learn, NumPy, Pandas, Matplotlib, Seaborn, NLTK, DBT
  • AWS Cloud · Lambda, EC2, S3, DynamoDB, CodeCommit, Elastic Beanstalk, EFS, RDS, API Gateway
  • Databases · SQL, NoSQL, DynamoDB, PostgreSQL, Snowflake
  • Soft Skills · Diligent, quick learner, proactive, collaborative, critical thinker

Selected publications

Google Scholar Lattes DBLP

2023

Automatic Extraction of Conversation Flows from Human Dialogues: Understanding Their Impact to Refine NLP Models

SN Computer Science. Research on automatically extracting conversation flows from human dialogues to refine natural language processing models.

DOI: 10.1007/s42979-023-02148-7

Curricular Transfer Learning for Sentence Encoded Tasks

arXiv. Study of curriculum learning for gradual adaptation between distributions in sentence-encoded tasks.

arXiv: 2308.01849

2022

MCCD: Generating Human Natural Language Conversational Datasets

ICEIS 2022. A methodology for generating multi-turn, multi-user conversational datasets from online forums.

DOI: 10.5220/0011077400003179

Textual Datasets For Portuguese-Brazilian Language Models

Dataset Showcase Workshop, SBC. Presentation of textual datasets for Brazilian Portuguese language models.

DOI: 10.5753/dsw.2022.224294

2021

EFIS: Ecological Fuel-consumption Intelligent System

DCOSS 2021

2020

RP-Sim: Radio Propagation Simulator

SBRC Companion 2020

DOI: 10.5753/sbrc_estendido.2020.12395

Projects

RP-Sim

Radio propagation simulator

An open-source environment for simulating wireless communication propagation, with support for propagation models, RSSI, BER, communication-area export, and an interactive demo.

Technologies: JavaScript, p5.js, Mappa.js.

Repository Demo

MCCD

Conversational methodology and datasets

A methodology for generating natural conversational datasets from online forums and other sources of human dialogue.

Repository

Miner-XenForo

Forum-mining tool

A tool for extracting and preparing conversational data from forums for use in natural language processing datasets.

Technologies: Python.

Repository

Portuguese NLP dataset

Datasets for natural language processing in Portuguese

A collection of new or adapted datasets for training Portuguese-language models.

Repository

GGA

Python library for genetic algorithms

An open-source library that simplifies the creation and customization of genetic algorithms in Python.

Repository

Msim

Event simulator

A simple event simulator written in Python.

Technologies: Python.

Repository

Satellite Image Processing

Satellite image processing

Processing Landsat 8 and Sentinel-2 satellite imagery to support farmers’ decision-making with color imagery and NDVI.

Technologies: AWS, Python, GDAL, Rasterio, Pillow.

SCO3, Servidor Clickideia Offline 3.0

FAPESP project

A project focused on automatically updating educational content through opportunistic networks in schools with limited or no internet access.

BV FAPESP

Awards and recognition

2022

Best Paper Award, ICEIS 2022

Paper: MCCD: Generating Human Natural Language Conversational Datasets.

2019

Second place in the Roberto Diéguez Galvão Award, SBPO 2019

Paper: Genetic Algorithm for the Maximum Coverage Location Problem Applied to Medical Emergency.

  • 2025: Professional recognition for creative and innovative initiative · Orion Sistemas Agrícolas.
  • 2016: Professional recognition for creative and innovative initiative · CYGNI AgroCiência.
  • 2014: Fourth place in the regional programming contest · Sociedade Brasileira de Computação.
  • 2013: First place in UNIVEM’s computing olympiad for high school students.