CareerPlanSign in

Mathematical Scientist for AI Safety Research

Montreal💼 Full-time🗓 2026-05-22 → 2026-09-26

Core

Theoretical research to develop a novel, safe-by-design AI system (Scientist AI) with probabilistic safety guarantees under finite data and compute constraints.

Role type

Principal theoretical researcher (AI safety)

Builds

Theoretical blueprint for safe AGI and corresponding safety guarantees

Domain

AI safety, machine learning theory, formal methods

Deliverable

production ML models

Required skills

formalization of natural-language intuition, abstract systems thinking, probabilistic machine learning theory, neural network analysis, language model analysis

Preferred skills

technical AI safety experience, conference travel

Technologies

next-token-prediction, DPO, RLHF, RLVR

Responsibilities

Explore new AI system architectures and safety mechanisms; analyze and refine formal requirements for safe deployment; write and disseminate scientific papers; guide research agenda to prioritize critical safety experiments; share knowledge to help design new ML algorithms

Seniority

Senior, hands-on IC

Sourced via greenhouse · Listed on CareerPlan, which tracks 70,000+ jobs from 20+ sources.