Open to thoughtful work
Let’s make reliable AI more useful.
I study calibration under safety alignment and ways to keep safe models useful. If your work overlaps with the areas below, I would be glad to hear from you.
Jakarta working internationally
Start with the kind of conversation.
A useful first message does not need to be long. It should make the problem and the next decision visible.
Research Collaboration
ResearchI welcome work on the alignment tax, calibration, honesty, abstention, and evaluation beyond English. I am especially interested in measuring the costs of safety training and where they fall.
The research question, available evidence or data, and the contribution you have in mind.
What happens nextIf it overlaps, I will suggest a concrete question or next working conversation.
Research & Applied Roles
RolesI am open to research and applied scientist roles worldwide, especially in alignment finetuning, calibration, evaluation, and reliable multimodal or multilingual systems.
The problem the team owns, the expected scope, location or working model, and likely timeline.
What happens nextI will reply with fit, availability, and any questions needed before a formal process.
Speaking
EventsI speak at conferences, universities, and panels about AI safety, trustworthy and multimodal AI, uncertainty quantification, and AI for underrepresented languages.
The audience, format, proposed topic, date, location, and whether travel or remote participation is expected.
What happens nextI will confirm topic fit, availability, and the format I can contribute most usefully.
Mentorship
GuidanceI mentor early-career researchers and engineers entering AI safety and trustworthy AI. This includes graduate applications and applied science careers, especially across Southeast Asia.
Your current stage, the decision you face, what you have tried, and the specific feedback you need.
What happens nextWhen capacity allows, I will offer a focused reply, resource, or short conversation.
