Customer-obsessed science

Time-series diagnostic reasoning is essential for many applications, yet existing solutions face a persistent gap: general reasoning large language models (GRLMs) possess strong reasoning skills but lack the domain-specific knowledge to understand complex time-series patterns. Conversely, fine-tuned time-series LLMs (TSLMs) understand these patterns but lack the capacity to generalize reasoning for more

As large language models (LLMs) continue to scale and new GPUs are released even more frequently, there is an increasing demand for LLM post-training in heterogeneous environments to fully leverage underutilized mid-range or previous-generation GPUs and alleviate the shortage of homogeneous high-end GPUs within a single availability zone. However, achieving high-performance reinforcement learning (RL) training

Large language models (LLMs) are increasingly deployed in decision-making tasks, where not only accuracy but also reliable confidence estimates are essential. Well-calibrated confidence enables downstream systems to decide when to trust a model and when to defer to fallback mechanisms. In this work, we conduct a systematic study of calibration in two widely used fine-tuning paradigms: supervised fine-tuning

Pretrained models are routinely extended with new vocabulary entries, for example new items in recommender systems and new tokens in large language models (LLMs). We show that jointly training old and new embeddings has a hidden failure mode: old-entry quality degrades while new entries are still learning. On sequential recommendation, old items overfit while new items improve, forcing premature early stopping

Pretrained time-series foundation models can condition on future-known covariates, but event covariates are often limited to binary indicators or sparse categorical labels. This is problematic for forecast-based anomaly detection: in financial transaction series, recurring holidays, promotions, and settlement cycles induce predictable shifts that event-unaware systems may flag as false positives. We propose