Cheng-Yuan (Ross) King
AI Safety & Evaluation · MSc Artificial Intelligence
MSc Artificial Intelligence candidate at the University of Sheffield. I build empirical AI safety systems — AI-agent evaluation, red-team testing, safety classifiers, RAG benchmarks, and reproducible Python workflows — with a focus on measuring model behaviour and failure modes.
Selected work
Selected technical projects
Representative projects in AI safety evaluation, LLM red-teaming, RAG benchmarking, and reproducible evaluation infrastructure. Full case studies on the projects page.
2026
Synthetic evaluation lab for AI-agent reliability: 350 golden cases, 60 red-team cases, RAG evaluation, safe refusal, safety classifiers, prevalence estimation, human-review simulation, mitigation impact, release gate reporting, FastAPI, OTel tracing, and CI.
- Python
- FastAPI
- Streamlit
- Pydantic
- pytest
- ruff
- Docker
- GitHub Actions
- GitHub Pages
2026
Reproducible benchmark measuring how published adversarial prompts perform against 2026-era LLMs and whether prompt-only defences move the needle — with cross-judge validation and bootstrap confidence intervals.
- Python
- Claude Sonnet 4.6
- Llama 3.1 8B
- Inspect AI
- GitHub Actions
- pytest
- ruff
- mypy
2026
End-to-end synthetic fintech data product — dbt metrics, CUPED A/B experimentation, activation model, geo-lift referral analysis, pricing intelligence, FastAPI service, and a full GCP deployment path with BigQuery, Cloud Run, and Cloud Monitoring.
- Python
- dbt
- DuckDB
- BigQuery
- Cloud Run
- FastAPI
- Streamlit
- Marimo
- scikit-learn
- synthetic control
- GitHub Actions
- GCP
Toolbox
What I work with
Pragmatic stack — pick the right tool, ship, measure, iterate.
- AI Safety & Eval
- RAG evaluation
- LLM evaluation
- Prompt engineering
- Structured extraction
- Red-team testing
- Guardrail checks
- Hugging Face Transformers
- Safe refusal evaluation
- Adversarial testing
- Safety classifier evaluation
- OpenTelemetry / OTel tracing
- AI & machine learning
- PyTorch
- scikit-learn
- PySpark
- Feature engineering
- Model evaluation
- Calibration
- A/B testing & CUPED
- Data & cloud
- dbt
- DuckDB
- BigQuery
- GCP
- Cloud Run
- Cloud Storage
- SQL
- PostgreSQL
- Engineering
- Python
- FastAPI
- Streamlit
- Pydantic
- Docker
- GitHub Actions
- Monitoring
- Analytics
- Synthetic control
- Causal inference
- Customer segmentation
- Dashboards
- pandas
- Statistical analysis
- Languages
- English (fluent)
- Mandarin (native)
- Japanese (JLPT N1)