5 sources
A high-signal person to follow for the part of alignment research that asks whether a model’s stated reasoning can actually be trusted and measured.
Researcher Profile
Editor reviewedSandipan Kundu
Alignment via AI feedback (Constitutional AI)
Robustness and evaluation researcher at Anthropic
Useful for the attack-and-evaluation side of alignment work, especially long-context jailbreak research and the measurement work that turns safety concerns into concrete tests.
Organizations
Anthropic
Labs
About This Page
This profile is meant to help you get oriented quickly: why this researcher matters, what to read first, and where to explore next.
Last reviewed
March 18, 2026
Known For
The ideas, systems, and research directions that make this person worth knowing.
01
Long-context jailbreak research
02
Reasoning-faithfulness measurement
03
Evaluation-heavy safety research
04
Alignment via AI feedback (Constitutional AI)
05
Constitutional AI: Harmlessness from AI Feedback
06
Training a Helpful and Harmless Assistant with Reinforcement Learning from Human Feedback
Start Here
Canonical papers, project pages, or repositories that anchor this profile.
Signature Works
Additional papers, projects, or repositories that help flesh out the profile.
Supporting Sources
Additional links that help verify and flesh out this profile.
Related Researchers
People worth exploring next because they share topics, labs, or source material with this profile.
5 sources
A high-signal page for anyone tracking whether model reasoning traces are actually trustworthy, not just fluent explanations pasted on after the fact.
5 sources
A strong person to follow for how Anthropic moved from assistant training into more explicit evaluation work around model behavior, red-teaming, and chain-of-thought faithfulness.
5 sources
A good person to follow for the part of alignment work that becomes concrete measurement: model-written tests, chain-of-thought faithfulness, and behavior-shaping methods that can actually be audited.
5 sources
Worth tracking for the newer evaluation thread at Anthropic, especially where failure-mode discovery and faithfulness measurement extend beyond the original RLHF papers.
4 sources
Important because he sits near the boundary between alignment theory and concrete failure-mode discovery, especially jailbreaks, preference training, and behavior evaluations.