About Me

I am was a Computer Science PhD student at Dalhousie University under the supervision of Dr. Evangelos Milios at the MALNIS lab. My research lies at the intersection of Deep Learning, Natural Language Processing and Information Retrieval. Before doing my PhD, I used to be a pure mathematician studying Differential Geometry and Ordinary Differential Equations, but after working for some years in Industry and Government, I decided it was time to study something that my family and friends could actually understand. This website is just a hub for my past and ongoing projects. Thanks for visiting!

Contact

Email: juan{.}ramirez{.}orta{at}dal{.}ca
Resume
LinkedIn
GitHub
Google Scholar

News

2023/11/30 - I passed my Thesis Defense! I am oficially a Doctor!
2023/10/13 - I passed my Thesis Proposal Defense!
2023/08/10 - My updated Overleaf Poster Template with the new colours and logo of Dalhousie University was approved!
2023/04/19 - My latest paper, QuOTeS: Query-Oriented Technical Summarization, was accepted at ICDAR 2023!
2022/11/15 - My team obtained a Global Finalist Honorable Mention in the NASA International Space Apps Challenge 2022!
2022/10/03 - My team won First Place in the SpaceApps Challenge by NASA and ShiftKey Labs!
2022/06/09 - My team won First Place in the second task of the DETESTS shared task at IberLEF 2022!
2022/02/04 - My team won the prize for the Best Presented Proposal at the Data Student Challenge from Volta and Research Nova Scotia!
2021/12/01 - My first serious paper, Post-OCR Document Correction with Large Ensembles of Character Sequence-to-Sequence Models, was accepted at AAAI 2022!
2021/11/15 - Attending my first GitHub Issue from one of my repositories!
2021/11/14 - I realized I obtained my first citation, my first GitHub Star and my first GitHub fork!
2021/10/28 - I passed my Research Aptitude Defense!
2021/03/15 - My very first paper, Unsupervised Document Summarization using Pre-trained Sentence Embeddings and Graph Centrality, was accepted at the Second Scholarly Document Processing Workshop at NAACL 2021!
2020/01/16 - I started my PhD!

Work Experience

Senior Data Scientist, Royal Bank of Canada

March 2024 - Present
Test, validate and document Artificial Intelligence models from all across the bank, focusing on the deployment and testing of Large Language Models.

Director, Halifax Chess Club (volunteering)

June 2022 - Present
Monitor the club media, deliver lectures, organize tournaments, look for sponsors and manage the finances of the club.

Data Scientist, Dalhousie University (contract full-time)

January 2024 - June 2024
Designed and implemented the front- and back-end of the Early Diagnosis Program in collaboration with the Nova Scotia Health Authority and developed the current implementation of the Explainable Hierarchical Binary Classifier.

Teaching Assistant, Dalhousie University (part-time)

April 2021 - December 2021
Elaborated fully-automated exam questions using R and Python to prevent cheating during COVID-19 in the courses of Theory of Computer Science and Foundations of Machine Learning.

Data Engineer, Banco Azteca

January 2019 - December 2019
Applied NLP to optimize the client data curation process of the bank resulting in an speed improvement of over 15%, and created automated reports of the customer databases using Python, R, SQL and Hadoop.

Senior Data Scientist, INFOTEC

May 2018 - January 2019
Developed and implemented an algorithm to index the publications of over two million scientists supported by the National Council of Science and Technology using NLP, Python and Linux. This algorithm is still currently in use by the National Archive of Science.

Graduate Statistics Professor, INFOTEC

January 2018 - January 2019
Developed teaching materials, delivered lectures and evaluated groups of over 30 students of the Master’s Degree in Data Science.

Data Scientist, INFOTEC

November 2016 - May 2018
Developed automated indicators to monitor 300,000 databases of the Mexican Open Data Portal to enable their use by Mexican citizens using R and Python.

Education

PhD in Computer Science, Dalhousie University

Area: Natural Language Processing and Machine Learning
Thesis: Local Methods for Document-level Natural Language Processing
Supervisor: Dr. Evangelos Milios
Jan 2020 - Nov 2024
GPA: 4.13/4.3

MSc in Mathematics, National Autonomous University of Mexico

Area: Topology and Complex Analysis
Thesis: Linear Differential Equations and Holomophic Vector Bundles
Supervisor: Dr. Laura Ortiz Bobadilla
Jan 2014 - Jan 2016
GPA: 9.4/10

BSc in Mathematics (Summa Cum Laude), National Autonomous University of Mexico

Area: Geometry and Ordinary Differential Equations
Thesis: Holomorphic Vector Bundles on Analytic Manifolds
Supervisor: Dr. Laura Ortiz Bobadilla
Aug 2009 - Jul 2013
GPA: 9.8/10

Publications

ROUGE-SciQFS: A ROUGE-based Method to Automatically Create Datasets for Scientific Query-Focused Summarization. COLING 2025.
QuOTeS: Query-Oriented Technical Summarization. ICDAR 2023.
MALNIS at IberLEF-2022 DETESTS Task: A Multi-Task Learning Approach for Low- Resource Detection of Racial Stereotypes in Spanish. IberLEF 2022.
Post-OCR Document Correction with Large Ensembles of Character Sequence-to-Sequence Models. AAAI 2022.
A Web Platform for Collaborative Semi-Automatic OCR Post-Processing. AGRANDA 2021.
Unsupervised Document Summarization using Pre-Trained Sentence Embeddings and Graph Centrality. Scholarly Document Processing Workshop at NAACL 2021.
Detection of Phonetically Similar Words in the Spanish of Central Mexico. Spanish Journal of Applied Linguistics 2020.

Awards

Global Finalist Honorable Mention, NASA International Space Apps Challenge, 2022
First Place, NASA Atlantic Canada Space Apps Challenge, 2022
First Place, DETESTS Shared Task-2 at IberLEF, 2022
Best Presented Proposal, ResearchNS Student Challenge, Canada, 2022
Gold Medal, Mexico City Amateur Chess Championship, 2018
Bronze Medal, VIII International Logic Olympiad, Mexico, 2011
Bronze Medal, II International Junior Science Olympiad, Indonesia, 2005

Open-Source Projects

Pytorch Beam Search, A lightweight implementation of Beam Search for sequence models in PyTorch
FIONA, A topically-aware search engine for the NASA Technical Report Server