
2 - 10 employees
Founded 2025
🤖 Artificial Intelligence
📚 Education
🏛️ Government
Artificial Intelligence • Education • Government
SASH is a Singapore-based organization building a global ecosystem for trustworthy and safe frontier AI. It bridges governments, academics and industry across East and West to advance practical, technical AI governance research (verification mechanisms, frontier safety, agent governance) and to operationalize that research through fellowships, training programs for policymakers and practitioners, events, briefings, and a co-working hub for AI governance and safety practitioners.
🔥 39 minutes ago
🇸🇬 Singapore – Remote
💵 $120k - $180k / year
⏰ Full Time
🟡 Mid-level
🟠 Senior
📚 Research Engineer
👻 Ghost score 3%
Where this job comes from
Posted on SASH's careers page ↗, not through a job board
First seen by Remote Rocketship 39 minutes ago
Still open on SASH's site. Checked 23 minutes ago.

2 - 10 employees
Founded 2025
🤖 Artificial Intelligence
📚 Education
🏛️ Government
Artificial Intelligence • Education • Government
SASH is a Singapore-based organization building a global ecosystem for trustworthy and safe frontier AI. It bridges governments, academics and industry across East and West to advance practical, technical AI governance research (verification mechanisms, frontier safety, agent governance) and to operationalize that research through fellowships, training programs for policymakers and practitioners, events, briefings, and a co-working hub for AI governance and safety practitioners.
• Design, implement, and run evaluations of misalignment and loss-of-control risks in frontier models • Build realistic agent environments, scaffolds, and tool integrations for studying behaviour under increasing autonomy • Develop infrastructure for long-horizon experiments involving many model calls, actions, tools, and environment states • Own experimental reliability and reproducibility across model access, sampling, configuration, environment management, logging, and analysis • Work with research scientists to turn open-ended questions into tractable experiments • Identify cases where apparent model behaviour is an artefact of the evaluation • Build tools for inspecting and analysing large collections of model trajectories and comparing behaviour across models and experimental conditions • Contribute experimental methodology, infrastructure, and technical analysis directly to published research
• Strong Python and general software-engineering skills • Experience working with language models, agentic systems, or model evaluations • Ability to turn an underspecified research question into a reliable experimental setup • Strong engineering practices around reproducibility, testing, observability, and data integrity • Ability to investigate unexpected results across both the model and surrounding infrastructure • Clear technical writing and communication skills • Comfort working in an early-stage environment where requirements and research directions evolve • Mandarin reading or writing ability is helpful but not required • Experience with AI safety or dangerous-capability evaluations is helpful but not required • Experience with evaluation frameworks such as Inspect is helpful but not required • Experience building agent scaffolds or tool-use environments is helpful but not required • Experience with distributed inference, large-scale API-based experimentation, Docker, Kubernetes, or related infrastructure is helpful but not required • Experience analysing large collections of model trajectories or transcripts is helpful but not required • Familiarity with frontier-model safety reports is helpful but not required
• Competitive benefits and leave policies • Remote work option • Flexible location: Singapore or remotely
Apply NowImprove your chances of getting an interview by checking your resume score before you apply.
🕒 November 12
Research Engineer building PII detection and anonymization systems for an AI training-data marketplace. Preserving privacy while retaining useful data for downstream model training.