Multilingual AI · Search & Retrieval (RAG)

Hao (Peter) Yu

Member of Technical Staff at Boson AI · Researcher affiliated with Mila

I work on multilingual AI and Search & Retrieval (RAG), from data and training to evaluation and production systems.

Résumé Google Scholar Email
800+
scholar citations
8
h-index
2
first-author ACL oral papers
Hao Yu standing by the ocean at sunset
Toronto / Montréal

Selected work

Research

All publications

Multilingual AI

Data, training, and evaluation for African and low-resource languages.

20 languages · 26B tokens · 5 base models

  1. AfriqueLLM results comparing base and adapted models across language groups ACL 2026 Oral

    FlagshipMultilingual language models

    AfriqueLLM: How Data Mixing and Model Architecture Impact Continued Pre-training for African Languages

    2026
    Hao Yu Tianyi Xu Michael A. Hedderich Wassim Hamidouche Syed Waqas Zamir David Ifeoluwa Adelani
    ACL 2026 Oral Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)
    Paper Models GitHub arXiv DOI
    Abstract

    AfriqueLLM is an open suite of language models adapted to 20 African languages through continued pre-training on 26B tokens. The release includes model weights, data recipes, training configurations, and evaluation commands; systematic experiments across five base models show that task-aligned data composition is the primary driver of downstream gains.

Also INJONGO · 16-language benchmark.

Search & Retrieval (RAG)

Evaluation and reasoning for RAG, agentic search, and knowledge graphs.

First author · CCF BigData · 2024

  1. Auepora evaluation targets for retrieval-augmented generation CCF BigData

    FlagshipSearch and retrieval

    Evaluation of Retrieval-Augmented Generation: A Survey

    2024
    Hao Yu Aoran Gan Kai Zhang Shiwei Tong Qi Liu Zhaofeng Liu
    CCF BigData Big Data: 12th CCF Conference, BigData 2024
    Paper DOI
    Abstract

    We introduce Auepora, a unified process for evaluating retrieval and generation in RAG systems. The survey compares metrics, datasets, and benchmarks across relevance, accuracy, and faithfulness.

Also Search-on-Graph · graph reasoning.

Current roles

Experience

2025 — present

Boson AI

Member of Technical Staff

Multimodal evaluation and memory for real-time agents.

2022 — present

Mila

Affiliated Researcher

Multilingual models, retrieval, reasoning, and evaluation.

2024 — 2026

McGill University

M.Sc. in Computer Science (Thesis)

Multilingual NLP for low-resource languages · GPA 4.0 · Supervised by David Ifeoluwa Adelani .

2020 — 2024

McGill University

B.Sc., Major in Software Engineering

Faculty of Science · Supervised by Reihaneh Rabbany .

Beyond current work

Open directions

Quantum Computing & Simulation

Differentiable quantum simulation and NISQ software.

TensorCircuit
World Models for Robotics

Predictive learning for planning and control.

Vision-Language-Action Models

Connecting perception, language, and embodied action.

3D Design & Rapid Prototyping

CAD, 3D printing, and fast hardware iteration.

Interested in collaborating? Email me

Writing

Notes

Read notes

脚踏实地，行稳致远。

Stay grounded. Move steadily. Go far.
Contacthao.yu2@mail.mcgill.ca