Anshuman Mishra

“Co-authored technical post on building AI agents from first principles, demonstrating how agent training reduces to prompt-model-environment-reward-gradient loop.”