ARTIFICIAL INTELLIGENCE RESEARCHER

Naomi J Sutcliffe de Moraes

Institute of Mathematics and Statistics
University of São Paulo, Brazil

Naomi Sutcliffe de Moraes
About me

I specialize in artificial intelligence. My recent and current research projects involve large language models, generative AI, machine learning, natural language processing, algorithm development and the application of these fields to engineering, education and public health.

Educational Background

Born and raised in the United States, I started my academic journey at the University of California, Los Angeles (UCLA). After my B.S. in Mechanical Engineering, I continued on at UCLA to complete an M.S. in Physics. After working in industry in Los Angeles as a computational physicist and software engineer, I moved to Brazil. I then did a PhD in Linguistics at the University of São Paulo (USP). I will soon finish a second PhD in Computer Science from USP.

My Experience

2022–

University of São Paulo (USP), Institute of Mathematics and Statistics

Researcher

Automated Argument Mining and Argument Quality Feedback Project

The problem: Develop a prototype application to provide human-like feedback on student essay argument structure.

My approach:

  • Compiled a corpus of student essays with human feedback, a total of 2.4 million words, using web scraping, pre-processing techniques and data augmentation.
  • Used prompt engineering techniques on an LLM to obtain feedback based on theoretical argument structure and argument quality frameworks.

Result: Ongoing

Diagonal Circular Buffer Matrix Project

The problem: The RLS-DCD adaptive filtering algorithm requires O(M2) operations, where M is the size of the autocorrelation matrix. The objective was to develop a new data structure and algorithm that would run in O(M) time.

Our approach: Proposed and implemented a novel data structure called the Diagonal Circular-Buffer (DCB) Matrix Format.

Result: The improved RLS-DCD algorithm ran up to 94% faster in Julia and C than the baseline in simulations. (See article below in the publications section.)

2016–2018

Universidade Federal do ABC (UFABC), Center for Mathematics, Computer Science and Cognition

Visiting Professor

  • Taught Single-variable calculus, Multi-variable calculus and Introduction to ordinary differential equations to undergraduate students.
  • Taught academic English classes to undergraduate and graduate students and colleagues.
  • Created my own temporary social media site for interactive class assignments.
  • Published research in two international academic journals. (See list of publications below.)
  • Presented at two international academic conferences.
Prior Experience

Los Angeles, California

Software Engineer and Computational Physicist

  • Wrote computer simulations and analyzed the performance of defense systems.
  • Wrote software controlling US/Canadian join radar system (C++, UNIX)
  • Coordinated code changes in global files used by a team of 50 developers.
  • Developed algorithms for laser tracking of missiles in flight (C)
  • Performed laser tracking experiments with small models in the lab and carried out data analysis of results for different tracking algorithms.

Selected Publications and Conference Presentations

N. J. Sutcliffe de Moraes, V. H. Nascimento, D. C. Vidal, C. A. Prete and Y. V. Zakharov, A faster RLS-DCD adaptive filtering algorithm, 2024 19th International Symposium on Wireless Communication Systems (ISWCS), Rio de Janeiro, Brazil, 2024, pp. 1-5, doi: 10.1109/ISWCS61526.2024.10639158.

J. Luiz Pastore Mello, N. J. Sutcliffe de Moraes, Periodic functions, real numbers and least common multiples: Is the sum of two periodic functions also periodic? The Australian Mathematics Educational Journal, v. 1, p. 37-41, 2019.

N. J. Sutcliffe de Moraes, Using Massive Open Online Courses in an English-for-Academic-Purposes, Flipped-Classroom Context. In: 7th International Second Language Pedagogies Conference, 2018, Victoria, BC, Canada. Program of the SLPC7. Victoria, BC, Canada: University of Victoria, 2018.

N. J. Sutcliffe de Moraes, A Sentence a Day – How to Engage University Students in an Online Daily Writing Task. Nouvelle Revue Synergies Canada, p. 1-12, 2018.

MORAES, N. J. S. Linguistic differences between legal and medical translation. In: Tenth International Conference of the Brazilian Studies Association, 2010, Brasília. Anais do X Congresso Internacional da Brasa, 2010. p. 1-12.

Selected Certificates

  • Prompt Engineering for ChatGPT, Vanderbilt University
  • Data Analysis and Statistical Inference, Duke University
  • Weights & Biases 101
  • Weights & Biases 101 – Weave
  • Object-Oriented Programming (Parts 1 and 2), University of São Paulo
  • Object-Oriented Data Structures in C++, University of Illinois
  • Ordered Data Structures (C++), University of Illinois
  • Unordered Data Structures (C++), University of Illinois
  • Python Basics, University of Michigan
  • Introduction to Scientific Computing, University of São Paulo (graduate-level course)
  • Aleatoire (Probability), École Polytechnique, France
  • Social and Economic Networks – Models and Analysis, Stanford University
  • Game Theory, Stanford University
  • Model Thinking, University of Michigan
  • Test of Chinese as a Foreign Language (TOCFL), Taiwanese Government Band A, Level 2

Citizenship

  • US Citizen
  • Right to work in Brazil

Links to Personal Profiles

GitHub

LinkedIn

CV Lattes