Eric Thayer Email & Phone Number

Boston, MA

Greater Boston Area

Cambridge, MA

Site features, back-end services and messaging infrastructure, dealer/user support tools, internationalization, system monitoring/alerting, data pipeline

Cambridge, MA

Bringing the good stuff from Vlingo to Nuance.

Cambridge, MA

Early hire in the engineering team. Designed and implemented highly scalable, responsive and reliable speech servers and voice platform components.

Waltham, MA

Managed the speech group. Created and evaluated UK English acoustic models for a shipping UK release. Created SphinxTrain acoustic model training infrastructure using Python. Participated as a developer in a consumer release of the reading tutor.

Maynard, MA

Lead engineer for voice conferencing in a team that developed a real-time collaboration product (Jabber/XMPP based) that integrated multiple chat clients with desktop sharing and teleconferencing over the web. Tasks included:- Developing a 500 phone line (20 T1) voice bridge (included voice conferencing, IVR menus, in/out dialing, conference recording).

Waltham, MA

· Developed the server side data synchronization component of a web-based mobile e-mail/contact/calendar application for J2ME phones. System provided bi-directional syncing of e-mail/contacts/calendar info. Application data was acquired using Microsoft Exchange via MAPI in addition to a relational database repository of downloaded e-mail from POP3.

Waltham, MA

· Responsible for improving the CMU SPHINX-3 acoustic model training system. Introduced variable topologies across phones, variable number of densities/state. Managed a group of two other acoustic modelers.· Ran and analyzed experimental results to achieve a 60% reduction in model size w/o loss of accuracy· Built a turn-key model build/test environment.

Pittsburgh, PA

· Developed the SPHINX-3 acoustic model training system that creates large vocabulary state-tied-mixture fully-continuous HMMs from a training corpus of spoken utterances. Used in many DARPA speech evaluations.· Contributed to CMU DARPA spoken language evaluations beginning in 1993. Trained the acoustic models used in the primary evaluations and assisted.

Pittsburgh, PA

· Integrated CMU Sphinx system and NCSA Mosaic browser to allow voice control over browsing.· Developed a speech compaction algorithm that shortens the duration of spoken utterances by the deletion of extraneous silence regions and noise regions (fricatives). Could be used as a pre-processing step to other speech compression algorithms.· Optimized the.

Pittsburgh, PA

Pittsburgh, PA

Consulting while at CMU· Developed acoustic models for Fifth Generation Systems that were used in a system deployed by US West in its Midwest directory assistance call centers. (http://www.fifthgen.com)· Ported the SPHINX ASR system for the Sun Green Project (spinoff of that project was that crazy language called Java)· Designed and implemented acoustic.