cv

Full Name	João Augusto Leite
Languages	Portuguese, English

• 2022-2026

  PhD - Computer Science

  University of Sheffield, UK

  • Specialising in disinformation mitigation using LLMs and fine-grained signals (e.g., credibility, persuasion).
  • Supervised by Prof. Dr. Carolina Scarton, Prof. Dr. Kalina Bontcheva, and Dr. Olesya Razuvayevskaya.

• 2021-2024

  Msc - Computer Science

  Universidade Federal de São Carlos, Brazil

  • Specialised in self-training and data augmentation methods for hate speech detection.
  • Supervised by Prof. Dr. Diego Silva.
  • Published a research paper and a thesis: "Noisy Self-Training with Data Augmentations for Offensive and Hate Speech Detection Tasks". Presented at RANLP 2023 in Varna, Bulgaria.

• 2017-2021

  Bsc - Computer Science

  Universidade Federal de São Carlos, Brazil

  • Participated in two research labs: MaLL with Prof. Dr. Estevam Hruschka and MIDAS with Prof. Dr. Diego Silva. Published two research papers.

• 2024 - Present

  Research Associate

  University of Sheffield, UK

  • Developing content verification tools for the Chinese language.

• 2023 - 2024

  Graduate Teaching Assistant

  University of Sheffield, UK

  • Prepared lab demonstrations and graded undergraduate work for the text processing module (COM3110).

• 2021 - 2022

  Data Scientist

  PicPay, Brazil

  • Designed and deployed information retrieval models for millions of users.
  • Conducted A/B testing for proof-of-concept projects.

• 2019 - 2021

  Machine Learning Engineer

  Birdie.ai, Brazil

  • Built NLP models for applications like NER, sentiment analysis, and ontology building.
  • Led labeling tasks for supervised learning.

• 2019

  Organiser

  Semana da Computação - UFSCar, Brazil

  • Semana da Computação is an annual event that brings together students and professionals from many fields of Computer Science.
  • I was a member of the organising committee and responsible for defining the event's schedule and contacting speakers.

• 2022

  EPSRC Doctoral Training Partnership (DTP) Scholarship

  UK Research and Innovation (UKRI)

  • Covers full international tuition fees, living expenses and provides a research support grant.

• 2021

  Best Paper Award

  Department of Computer Science, Universidade Federal de São Carlos

  • My paper "Toxic language detection in social media for Brazilian Portuguese: New dataset and multilingual analysis" was awarded the best paper published in the computer science department in 2021.

For a complete list of publications, please visit my Google Scholar profile or the publications page.

• Applications

  • AI for social good, content verification, disinformation and hate speech mitigation, assessment of credibility and persuasiveness.

• Responsible AI

  • Fairness, transparency, alignment, adversarial robustness, interpretability.

• Natural Language Processing

  • Language modelling, agentic AI, prompting strategies, learning from explanations, surface-form competition, multilingual settings and cross-domain adaptation.

• Semi-supervised and Unsupervised Learning

  • Self-training, weak supervision, data augmentation, contrastive learning, zero and few-shot learning.

• Programming Languages

  • Python, C

• Machine Learning Frameworks

  • PyTorch, HuggingFace, Scikit-learn, spaCy, NLTK

• ML Ops

  • Git, PyTorch Lightning, Weights & Biases, DVC, Databricks, AWS

• Other

  • Unix shell scripting, Web scraping, Slurm