May 2022 – Apr 2024 · 1 yr 11 mos

May 2019 – Apr 2022 · 2 yrs 11 mos · Bengaluru Area, India

• Acquired by Samsung.
• Developed conversational AI tech stack for Indian languages: spoken language detection, speaker diarization, speaker verification, automatic speech recognition (ASR) and neural speech synthesis (TTS).

Jul 2018 – Nov 2018 · 4 mos · Redmond, Washington

Jun 2017 – Sep 2017 · 3 mos · San Francisco Bay Area

• Spatial audio visualization for VR applications

May 2016 – Apr 2019 · 2 yrs 11 mos · Tampere Area, Finland

• Worked on computational auditory scene analysis (CASA) using machine learning techniques
• Specifically on 3D sound event localization and detection in real life multichannel audio using machine learning methods.

May 2015 – Apr 2016 · 11 mos · Bengaluru Area, India

• Worked on an audio fingerprinting solution for unintentional recording.
• Applications like Shazam and Sound-hound work on intentionally collected user recordings. This problem becomes more complicated when the recordings are not intentional. Such non-intentional recordings are used for profiling mobile users based on their media consumption and further deliver the right ads.

Jun 2012 – May 2015 · 2 yrs 11 mos · Mumbai Area, India

• Worked on applications involving music information retrieval, speech and audio processing.
• Worked extensively on machine learning techniques for signal processing.

Sep 2011 – Jun 2012 · 9 mos · Tampere Area, Finland

• External Researcher at Nokia Research Center, Tampere
• Indoor smart-phone localization
• Developed a multi-device localization algorithm using ambient speech signals.
• Theory for asynchronous microphones localization using surrounding continuous sounds such as speech was established.
• A noise robust time difference of arrival (TDOA) technique was used to calculate the individual distances; and further mapping was done using multi-dimensional scaling (MDS) algorithm.

Feb 2011 – Sep 2011 · 7 mos · Tampere Area, Finland

• Technique to estimate reflection coefficient of room surfaces using audio impulse response was developed.
• Usage of such reflection coefficient’s in estimating room geometry from room impulse response was studied.

May 2009 – Jul 2009 · 2 mos · Bengaluru Area, India

• Embedded Systems R & D Group
• Worked on advertisement detection in cookery shows using audio features for a video summarization module.