I joined IBM Research as a fresh PhD in 1974 and spent many years working on Speech Recognition, eventually managing a group in the Software Division developing the IBM ViaVoice speech recognition product. Since 2006 I have been back in Research as part of a team developing Apache UIMA. Originally an internal IBM Research project, Apache UIMA is an open source, highly scalable, multimodal framework used to create applications doing deep analysis of natural human language text and speech.
Recently I have been responsible for the system that interfaces the Watson/DeepQA system to the Jeopardy! game computers. It manages the signals to all of Watson's appendages: the DeepQA brain, the hand that presses the buzzer, the voice that gives the answers, and the face displayed behind the podium.