2. realtime data context, for AI LLM that need fresh data context to do a better job, timeplus can be used to process and service those realtime data for it
3. realtime inference, in those cases where user want to do realtime inference, leveraging those UDF, user can combine the realtime data with model inference in one single box, greatly simplify the overall ML architecture