Engineering
Building the tools and infrastructure that make AI systems measurable, testable and trustworthy.
Sound engineering is the foundation of trustworthy AI. We build open evaluation tooling, test harnesses and safety infrastructure so that claims about AI systems can be checked rather than simply believed. Our engineering work serves the whole ASEAN research community and is open for others to use and build upon.
Focus areas
Evaluation tooling
We develop open-source platforms for running evaluations, so researchers and practitioners can test model capabilities against shared, reproducible benchmarks.
Test harnesses
We maintain harnesses for safety and security testing, including automated probes for known risk categories.
Safety infrastructure
We build infrastructure that helps organisations monitor deployed systems and respond when behaviour drifts from expectations.
Open-source tooling
Wherever possible we release our tooling openly so the wider community can inspect, extend and trust it.
Instruments
Beneath and beyond, now and next.
liwaiwai labs is an independent AI research lab for the ASEAN region. We work across research, engineering, assurance and deployment, serving both industry and community.
Explore our frameworks