Member of Technical Staff - Research (AI Security)

General Analysis is an AI safety and security research lab building guardrails, red-teaming frameworks, evaluations and runtime protection for AI and coding agents. As a Research Engineer you'll own the training pipeline, from data generation and environment design to evaluations, using reinforcement learning on distributed systems.

What you'll do

• Post-train models using reinforcement learning, across the full training pipeline.
• Design data generation, RL environments and evaluations.
• Solve engineering challenges in distributed systems and low signal-to-noise settings.
• Make algorithmic improvements and apply research to real-world AI security problems.

Requirements

• 0–5 years as a researcher or software engineer in AI security, benchmarking or RL.
• Comfortable with PyTorch and RL / post-training infrastructure (e.g. Tinker, verl).
• Academic or industry ML/RL experience; publications at NeurIPS, ICML, ICLR or similar are a plus.
• High agency, strong technical acumen and clear thinking.

Compensation: $250K – $450K + 0.5–1.5% equity. 5 days in office in San Francisco. Open to H-1B transfer, TN and STEM OPT.