Amazon Technologies, Inc.

Asem Rustum

Rahul Gupta

Matthew Scott Garman

Derek Solomon Pai

Sravan Babu Bodapati

Jin Dong Kim

Mohit Gupta

Joshua Marc Burgin

Katrin Kirchhoff

Christoph Saalfeld

Douglas Cotton Kurtz

Joby Sky Lafky

Stephen Alden Elliott

Varun Sembium Varadarajan

Charles Eric Dannaker

Pu Paul Zhao

Aniruddha Shriram Daptardar

Vivek Govindan

Julian E S Salazar

Calvin Xuan Li

Sunitha Muthireddy

Adam Bordianu

Sippakorn Tansutthiwess

Christian Egon Berkhoff Dossow

Himalay Mohanlal Joriwal

Sai Madhuri Bhavirisetty

Abhinav Kumar

Jason Daniel Ganzhorn

Orestis Lykouropoulos

Lawrence Wheeless

Akshay Nalwaya

Aditya Subhash Jadhav

Liangwei Guo

Yibin Wang

K P N V D S Siva Rama

A transcription service may receive a request from a developer to build a custom speech-to-text model for a specific domain of speech. The custom speech-to-text model for the specific domain may replace a general speech-to-text model or add to a set of one or more speech-to-text models available for transcribing speech. The transcription service may receive a training data and instructions representing tasks. The transcription service may determine respective schedules for executing the instructions based at least in part on dependencies between the tasks. The transcription service may execute the instructions according to the respective schedules to train a speech-to-text model for a specific domain using the training data set. The transcription service may deploy the trained speech-to-text model as part of a network-accessible service for an end user to convert audio in the specific domain into texts.

Model training system for custom speech-to-text models

Systems and methods permit customers to purchase resource instances on a spot purchasing mode basis, which permits a customer to specify a maximum price that the customer is willing to pay for one or more resource instances (e.g., virtual machine instances running on physical servers). Further, the resource instances purchased under the spot purchasing mode can be guaranteed to run for at least a specified period of time, even if the current variable spot cost value is exceeds the original bid price. Such instances run for defined blocks of time which may be selectable by the customer, thereby providing the customer some level of assurance that the instances will not terminate prematurely due to spot cost value fluctuations. Instances are assigned into a spot block instance pool from other types of pool based on the estimated excess capacity of such other types of pools. The estimate of excess capacity is based on statistical analysis of the usage of the instances.

Interruptible resource instances for block time periods

In a computing environment, a request to fulfill a computational task and a constraint for fulfilling the computational task is received from an entity. The request is independent of the quantity or type of resource slot to fulfill the computation task. The quantity or type of resource slots sufficient to fulfill the request is determined in accordance with a constraint. The resource slots are associated with the entity and allocated a plurality of geographically separate computing environments. Master resource slots are allocated in the geographically separate computing environments based on the criterion and the quantity or type of resource slots. The master resource slots can be reallocated based on changes to the allocated resources slots.

Distributed resource allocation

A computing device generates a plurality of visual indications of connections made between customers and service representatives by an automated dispatcher system. The computing device further provides animations of the plurality of visual indications moving on a map between customer locations associated with the customers and service representative locations associated with the service representatives.

Contact routing visualization system

Growing community of inventors

Renton, WA, United States of America

Asem Rustum