Travis Grizzel

Snoqualmie, Washington, United States Contact Info
3K followers 500+ connections

Join to view profile

About

I am a technologist; an architect, an engineer, a hacker, and a technical advisor. I am a…

Activity

Join now to see all activity

Experience & Education

  • Oracle

View Travis’s full experience

See their title, tenure and more.

or

By clicking Continue to join or sign in, you agree to LinkedIn’s User Agreement, Privacy Policy, and Cookie Policy.

Licenses & Certifications

Patents

  • NON-SPEECH INPUT TO SPEECH PROCESSING SYSTEM

    Issued 11990120

    This patent relates to a wearable device that a user can wake for speech processing with movement, such as a by lifting your wrist/hand towards your mouth. The wearable device has a motion sensor and uses a model to determine that the device has moved with a body part about a first axis and that the movement corresponds to a wake gesture for speech processing. The device captures audio and, because of the wake gesture, sends the audio for speech processing.

    See patent
  • NON-SPEECH INPUT TO SPEECH PROCESSING SYSTEM

    Issued US 10,692,485

    This patent relates to using motion data to interact with a natural understanding system. For example, the motion sensor data (from, e.g., gyroscope, accelerometer or proximity sensor) is associated with audio data from a spoken utterance. The user device sends the motion data as metadata with audio data representing an utterance to a natural understanding system, like Alexa. The system then performs speech processing on the audio data to get NLU data and uses both the NLU data and the motion…

    This patent relates to using motion data to interact with a natural understanding system. For example, the motion sensor data (from, e.g., gyroscope, accelerometer or proximity sensor) is associated with audio data from a spoken utterance. The user device sends the motion data as metadata with audio data representing an utterance to a natural understanding system, like Alexa. The system then performs speech processing on the audio data to get NLU data and uses both the NLU data and the motion data to determine what command is intended

    See patent
  • NON-SPEECH INPUT TO SPEECH PROCESSING SYSTEM

    Issued US 10,692,489

    This patent relates to using motion data to interact with a natural understanding system. For example, earbuds or eyeglasses may capture motion data representing a user's head nodding or shaking in response to a system prompt for a "yes" or "no" answer. As such, the motion data can be used to supplement or replace audio data when a voice user interface, such as Alexa, asks a question. The system thus receives audio data from the wearable device and performs ASR on the audio data. The system…

    This patent relates to using motion data to interact with a natural understanding system. For example, earbuds or eyeglasses may capture motion data representing a user's head nodding or shaking in response to a system prompt for a "yes" or "no" answer. As such, the motion data can be used to supplement or replace audio data when a voice user interface, such as Alexa, asks a question. The system thus receives audio data from the wearable device and performs ASR on the audio data. The system determines a command potentially matching the resulting text but also determines that further user input is needed. In response, motion data is received from the device, which represents the device being rotated about an axis (e.g., a head nod or head shake to represent yes or no). The system figures out that the motion data goes with the command and is sufficient to use to process the command. This allows the natural understanding system to take physical gestures as inputs meaning a user can do things like shake her head while wearing headphones to signal “no” in response a TTS prompt. Thus the system can engage in back and forth communication with a user even in noisy situations where processing audio may be difficult.

    See patent
  • NON-SPEECH INPUT TO SPEECH PROCESSING SYSTEM

    Issued US 10,515,623

    This patent relates to enabling a voice user interface system, such as Alexa, to receive commands from a user in noisy environments from a wearable Alexa-enabled device (Echo Buds) having motion sensors to detect a commands (head nodding/shaking). For example, after detecting the audio signal quality is below an acceptable threshold, the system allows the user's movements to be used as a substitute for speech data (e.g., nodding or shaking of head instead of saying “yes” or “no”). The system…

    This patent relates to enabling a voice user interface system, such as Alexa, to receive commands from a user in noisy environments from a wearable Alexa-enabled device (Echo Buds) having motion sensors to detect a commands (head nodding/shaking). For example, after detecting the audio signal quality is below an acceptable threshold, the system allows the user's movements to be used as a substitute for speech data (e.g., nodding or shaking of head instead of saying “yes” or “no”). The system receives sensor data from two different sensors and figure out that the sensor data indicates the wearable device is moving about a vertical or longitudinal axis, meaning the user is performing a head motion such as nodding or shaking her head. The system can then execute a command based on the head motion, thus allowing the head motion to substitute for a spoken command or wakeword in poor audio conditions. These operations can be performed by the back-end cloud system or by the wearable itself.

    See patent

Projects

  • Moneyball

    Project Moneyball is an internal initiative aimed at leveraging consistent data points across the organization to make data-driven decisions about headcount needs, alignment and timing. This project helped the SA organization understand the time to ramp an SA to "revenue productive", and enabled the SA organization to make hire ahead decisions.

  • Master Builder Program

    The Master Builder Program (MBP) is an internal SA-enablement initiative that aims to better prepare newly hired AWS SA resources for the type of work they'll be doing with our customers. MBP replaces the Americas SA organization's previous on-boarding initiative, and mimics realistic customer interactions that range from project inception to implementation, while shortening the overall timeline to ramp new SAs from 90 to 60 days.

  • AWS Trusted Advisor

    Core member of the AWS Trusted Advisor team, and helped shape this product from early inception to launch, and beyond.

    See project

More activity by Travis

View Travis’ full profile

  • See who you know in common
  • Get introduced
  • Contact Travis directly
Join to view full profile

Other similar profiles

Explore collaborative articles

We’re unlocking community knowledge in a new way. Experts add insights directly into each article, started with the help of AI.

Explore More

Add new skills with these courses