Boxi YuEmail

Research · products · teaching

Human-centric AI and software.

I am Boxi Yu, a software engineering researcher at Lero. I study how to test and evaluate coding agents and AI systems so that people can decide when to trust them, then turn those lessons into tools people can use and courses people can learn from.

Explore the research See the products

Google ScholarGitHubXEmail

Portrait of Boxi Yu
Boxi YuSenior Research Fellow at Lero, Ireland University of Limerick

Research

All research
How do we establish that software and AI behave correctly?

Evaluating coding agents

Strengthening tests and examining the reliability of benchmark results.

Constructing test oracles

Using relations between programs and executions to detect inconsistencies.

Testing AI systems

Controlled transformations reveal errors in vision and language systems.

AIOps & empirical evaluation

LightAD, OpenRCA 2.0, and CLEANet examine model efficiency, causal diagnosis, and robustness to contaminated data.

Connected research themes, with links to individual works. The grouping describes shared questions rather than a chronological sequence.

Co-founder

Products

All products

Alongside my research I co-founded OneAI and Pacific Studio, where we build AI products that keep people in the decision seat.

Co-founder · In use · AI education

ONEAI

Assessment Studio

Turn a learning objective into a question set, rubric, and student-level feedback loop. Teachers stay in the decision seat; AI handles the first pass.

In use at four schools in Guangzhou and Shenzhen, including olympiad training in mathematics, physics, chemistry, biology, and informatics at The Affiliated High School of South China Normal University and Guangzhou University Affiliated Middle School.

GenerateReviewExplain
See the workflow
Demo · student guidance, assessment generation, teacher review · 1:33

Co-founder · build, prove, understand

Pacific Studio

See the studio

Three products from Pacific Studio, which I co-founded, show how AI can support the software lifecycle: ship the work, check the proof, and read the ideas that shape it.

01Released

Build

OpenASE

Issue-driven orchestration for the full software lifecycle, with Claude Code, Codex, and Gemini CLI working across design, development, testing, and deployment.

Open source · Go · multi-agentView on GitHub
02Released

Prove

VeriFormal

AI-assisted formal verification that turns plain code into machine-checked proof goals, with an offline workflow for C, Java, Dafny, and Rust.

Formal methods · no egressStart verifying
03Beta

Understand

KeyLight

An interactive paper reader that combines layout-aware parsing, sentence-level highlights, bilingual reading, and answers that point back to the source.

PDF reader · research workflowTry the beta

Teaching · open course · Fall 2026

Agentic Software Engineering

Course & teaching Open course site

Agents generate. You verify, decide and explain.

A practical twelve-week course for building with coding agents while keeping specifications, tests, review, security, and judgment in the loop.

I have also mentored ten students, eight of whom are now Ph.D. students. See where they went.

News & writing

All news Blog