Abstract: Systems and methods are described herein for an efficient storage layout of recorded content associated with a particular user. Content segments, unique to the user and encoded/transcoded at different bit rates, may be stored/partitioned based on the likelihood of a particular bit rate version of content being requested by the user and a duration of playback for the content segment. Content that is more frequently requested may be concatenated in a single storage location on more high performance hardware. Further, content that is played back for a longer duration of playback may also be grouped together and stored on more high performance hardware. Content that is more likely to be played for only a short time may be stored within a plurality of storage containers.

Abstract: A method for managing conversation history in an AI system is disclosed. A user message is received and full conversation history payload is retrieved from a memory service. The full conversation history payload includes one or more messages, summaries, or moments. A prompt budget is dynamically allocated based on an available context window. The prompt budget determines token allocations for the messages, summaries, or moments from the full conversation history payload. A prompt is assembled for a language model by selecting conversation history elements to fit within the allocated prompt budget. The selecting balances between recent verbatim messages, summarized content of older messages, and relevant older moments. The prompt budget is iteratively adjusted and the prompt is iteratively reassembled as new information is added from one or more language model outputs or tool calls, while maintaining the full conversation history payload in the memory service.

Abstract: Methods and apparatus are disclosed for efficient storage and retrieval of content items, such as recorded content items of a cloud DVR system or other system storing content items.

Abstract: Disclosed are systems and methods for implementing distributed in-memory timer queues. A plurality of queues is distributed across a plurality of nodes. The plurality of queues includes a first queue. The first queue corresponds to current timer expirations for a tenant, time bucket, and shard. A sorted set includes minute references to unacknowledged timers for a tenant and shard. A plurality of processes is created. Each of the plurality of processes is configured to be a unique consumer of the distributed plurality of queues. A first consumer group and a second consumer group are created within each of the plurality of processes. The first consumer group corresponds to the first queue. The second consumer group corresponds to the sorted set. The plurality of processes is used to proportionally consume the first queue with the first consumer group and the sorted set with the second consumer group.

Abstract: Disclosed are systems and methods for implementing timers as a service. A request to create a timer is received. The request corresponds to a wait state condition associated with an event in a data stream. The request specifies an endpoint for a payload associated with an expiration of the timer. The timer is created based on the wait state condition. The creating of the timer includes publishing information pertaining to the timer in a database. The timer is consumed. The consuming of the timer includes processing the information pertaining to the timer. Based on the expiration of the timer, the payload associated with the expiration of the timer is sent to the specified endpoint.

Abstract: Methods and systems are described for storing content. One or more sections of the stored content may be subdivided into user-specific portions and common portions. The user-specific portions may be stored in user-specific storage associated with corresponding accounts of a plurality of user accounts. The common portions may be stored in common storage associated with the plurality of user accounts. User accounts may request access to the stored content. One or more caching criteria may be used to determine whether to cache any common portions of the stored content. Requests for the content may be processed using the common portions stored in the cache storage.

Abstract: A first server device may store a first copy of a content item, and a second server device may store a second copy of the content item. A computing device may receive a request for the content item from a user device. The computing device may determine a user identifier associated with the user device and send, to the first and second server devices, the user identifier and an indication of the content item. The computing device may determine, based on a response by the first server device, at least one of a quality score or a recording score associated with the first copy of the content item, and the computing device may send, to the user device and based on the at least one of the first quality score or the first recording score, a portion of the first copy of the content item.

Abstract: Disclosed are systems and methods for implementing distributed in-memory timer queues. A plurality of queues is distributed across a plurality of nodes. The plurality of queues includes a first queue. The first queue corresponds to current timer expirations for a tenant, time bucket, and shard. A sorted set includes minute references to unacknowledged timers for a tenant and shard. A plurality of processes is created. Each of the plurality of processes is configured to be a unique consumer of the distributed plurality of queues. A first consumer group and a second consumer group are created within each of the plurality of processes. The first consumer group corresponds to the first queue. The second consumer group corresponds to the sorted set. The plurality of processes is used to proportionally consume the first queue with the first consumer group and the sorted set with the second consumer group.

Abstract: Methods and systems are described for storing content. One or more sections of the stored content may be subdivided into user-specific portions and common portions. The user-specific portions may be stored in user-specific storage associated with corresponding accounts of a plurality of user accounts. The common portions may be stored in common storage associated with the plurality of user accounts. User accounts may request access to the stored content. One or more caching criteria may be used to determine whether to cache any common portions of the stored content. Requests for the content may be processed using the common portions stored in the cache storage.

Abstract: Disclosed are systems and methods for implementing timers as a service. A request to create a timer is received. The request corresponds to a wait state condition associated with an event in a data stream. The request specifies an endpoint for a payload associated with an expiration of the timer. The timer is created based on the wait state condition. The creating of the timer includes publishing information pertaining to the timer in a database. The timer is consumed. The consuming of the timer includes processing the information pertaining to the timer. Based on the expiration of the timer, the payload associated with the expiration of the timer is sent to the specified endpoint.

Abstract: A first server device may store a first copy of a content item, and a second server device may store a second copy of the content item. A computing device may receive a request for the content item from a user device. The computing device may determine a user identifier associated with the user device and send, to the first and second server devices, the user identifier and an indication of the content item. The computing device may determine, based on a response by the first server device, at least one of a quality score or a recording score associated with the first copy of the content item, and the computing device may send, to the user device and based on the at least one of the first quality score or the first recording score, a portion of the first copy of the content item.

Abstract: Disclosed are systems and methods for implementing distributed in-memory timer queues. A plurality of queues is distributed across a plurality of nodes. The plurality of queues includes a first queue. The first queue corresponds to current timer expirations for a tenant, time bucket, and shard. A sorted set includes minute references to unacknowledged timers for a tenant and shard. A plurality of processes is created. Each of the plurality of processes is configured to be a unique consumer of the distributed plurality of queues. A first consumer group and a second consumer group are created within each of the plurality of processes. The first consumer group corresponds to the first queue. The second consumer group corresponds to the sorted set. The plurality of processes is used to proportionally consume the first queue with the first consumer group and the sorted set with the second consumer group.

Abstract: Disclosed are systems and methods for implementing timers as a service. A request to create a timer is received. The request corresponds to a wait state condition associated with an event in a data stream. The request specifies an endpoint for a payload associated with an expiration of the timer. The timer is created based on the wait state condition. The creating of the timer includes publishing information pertaining to the timer in a database. The timer is consumed. The consuming of the timer includes processing the information pertaining to the timer. Based on the expiration of the timer, the payload associated with the expiration of the timer is sent to the specified endpoint.

Abstract: A computing device may receive a request to store a content segment(s). The request may include an estimated storage size. The computing device may determine a plurality of storage devices based on an estimated capacity of each of the storage devices to store the content segment(s). The computing device may cause each of the storage devices to allocate storage for the content segment(s). The allocated storage may be tracked and/or retrieved based on location metadata indicative of the allocated storage that may be stored at a cache and/or metadata repository.

Abstract: Methods and apparatus are disclosed for efficient storage and retrieval of content items, such as recorded content items of a cloud DVR system or other system storing content items.

Abstract: A computer-implemented method for generating optimized user journeys leveraging artificial intelligence is disclosed. The method includes receiving a user text prompt describing a desired journey objective and extracting context data for multiple users from a customer data platform. This context data encompasses attributes, events, predicted traits and audience memberships. A journey generation prompt is constructed by combining the received user prompt and extracted context data. This prompt is input into a machine learning model which processes the prompt to produce a user journey definition comprising interconnected nodes representing journey phases like audiences, waits and messages. The journey system validates the definition, modifying nodes to conform to predefined schema rules. The validated journey is output to campaign orchestration systems for execution across customer touchpoints. As journeys run, engagement data is collected for retraining models to improve journey performance over time.

Abstract: Methods and systems for providing supplemental data are disclosed. An exemplary method can comprise determining content currently being consumed by a user, determining one or more user preferences for display of supplemental data, retrieving supplemental data according to the one or more user preferences, and providing the supplemental data to the user concurrently with the content currently being consumed.

Abstract: Methods and systems are described for storing content that match topics of interest selected by a user or an automated process. Audio information associated with the content can be extracted, parsed, and grouped into topics. Incoming content with audio information that matches the topics of interest selected can be stored and made available to the user for later playback.

Abstract: Systems and methods are described herein for an efficient storage layout of recorded content associated with a particular user. Content segments, unique to the user and encoded/transcoded at different bit rates, may be stored/partitioned based on the likelihood of a particular bit rate version of content being requested by the user and a duration of playback for the content segment. Content that is more frequently requested may be concatenated in a single storage location on more high performance hardware. Further, content that is played back for a longer duration of playback may also be grouped together and stored on more high performance hardware. Content that is more likely to be played for only a short time may be stored within a plurality of storage containers.

Abstract: A computing device may receive a request to store a content segment(s). The request may include an estimated storage size. The computing device may determine a plurality of storage devices based on an estimated capacity of each of the storage devices to store the content segment(s). The computing device may cause each of the storage devices to allocate storage for the content segment(s). The allocated storage may be tracked and/or retrieved based on location metadata indicative of the allocated storage that may be stored at a cache and/or metadata repository.