From the source
- Aionic Labs' OpenTSLM, an open source family of time-series language models, is now available in
Amazon SageMaker JumpStart, letting organizations deploy it inside their own AWS environment.
- The models read operational measurements such as equipment sensor data and supply and demand alongside text context like maintenance records, market news and an engineer's observations, so systems can flag emerging problems in recent measurements.
- Amazon says general-purpose language models like Claude or ChatGPT can analyze data, but processing long histories across hundreds of operational metrics as text or images consumes large numbers of tokens and raises cost.
- TimeNet, the shared data infrastructure for time-series data, provides one format and API to find, build and load the datasets used to train and evaluate these models.
- Nexthink and Aionic Labs are applying the approach to proactive IT support, where Spark, Nexthink's personal IT agent, uses device telemetry and context to produce a state summary, a recommended action and an expected outcome.
- In an illustrative low-disk example, the model links declining free space, a rising disk queue and more frequent application hangs to symptoms such as a slow computer, and Spark runs an IT-approved cleanup after the employee confirms.






