Customer-obsessed science

At Amazon Prime Video, we face the critical operational challenge of managing code deployments during live events and rapid feature releases without causing service outages. Current change control approaches use blanket deployment freezes that block all changes regardless of risk, creating significant developer toil. While prior re-search has explored risky change predictors, these rely on developer-specific

Recent work identified Super Weights, individual parameters whose removal degrades model performance by orders of magnitude. We show that this degradation due to pruning Super Weights does not universally apply to all LLMs. Furthermore, if these parameters are so important, Super Weight-aware training should be effective. We show the opposite. Training Super Weights in isolation (100 to 8,192 parameters

Reward hacking arises when a model improves a proxy reward by exploiting shortcuts rather than solving the intended task. We study this failure mode through the geometry of reinforcement learning updates in language models and argue that hacking emerges when optimization drifts away from a stable low-dimensional learning trajectory. We analyze this drift through dominant singular directions of parameter

Large Vision–Language Models (LVLMs) increasingly rely on retrieval to answer knowledge-intensive multimodal questions. Existing benchmarks overlook conflicts between visual and textual evidence and the importance of generating deflections (e.g., 'Sorry, I cannot answer...') when retrieved knowledge is incomplete. These benchmarks also suffer from rapid obsolescence, as growing LVLM training sets allow

E-commerce stores face an evolving challenge in detecting fraudulent business registrations, as sophisticated actors continuously adapt their techniques to create deceptive accounts, rendering conventional post-registration measures increasingly inadequate. Realtime detection at the point of registration is further complicated by cold-start constraints and strict sub-second latency requirements. This paper