Jayakumar Singaram
Dr. Jayakumar Singaram is a veteran in semiconductor electronics and a Strategic consultant to companies such as Mistral Solutions, Bangalore, Bilva infra, Bangalore, SunPlus Software, Bangalore, and Apollo Tyre's, Chennai, among others. He drives deployment ready solutions using AI in IoT Edge/Node. He began his career working in Hindustan Aeronautics Bangalore, followed by Cranes InfoTech and Mistral Solution before launching a start-up world by steering brands like Epigon Media Technologies, Rinanu Semiconductor, and many more. He had been one of the core team members and also led the design and development of "Networked Karaoke Machine'' (which enabled TAITO Corp. Japan to become No. 2 in the world of Networked Karaoke).
Another major milestone has been his contribution to the development of Satellite Radio Receiver going by the brand name of ‘WorldSpace’. He is working on Deep Learning applications by using IBM Watson services. He has worked on key projects such as design and development of the Karaoke Machine (a project in collaboration with Analog Devices and MIT Media Lab) for TAITO Corp. He also worked on Satellite Radio Receiver for WorldSpace broadcast Satellites (and DAB Radio Receivers), by using low-cost Digital Signal Processors from Analog Devices.
He has a B.Tech in Aeronautical engineering from Madras Institute of Technology and also holds an undergraduate degree in Mathematics from Madras University. He has Masters and Doctorate degree in Systems and Control Engineering from Indian Institute of Technology, Bombay and a Research Fellow at KU Leuven University. As a Doctoral Thesis problem, he had worked on “Simultaneous Stabilization of Feedback Systems”.
He had worked as a Board Member and Dean of School of Computing Science and Engineering at Periyar Maniammai Institute of Science and Technology, Thanjavur, India, from 2007 to 2014. He was trustee of School of Music and Dance, Indiranagar-Bangalore, during 2009 to 2016. Link www.jkuse.com provides more details about him. He would like to thank the founders of Mistral Solutions, Anees Ahmed and Rajeev Ramachandra. And also he likes to thank his family members, Dr Suneetha Rao, Dr Vaidhehi and Niranjan Kumar for their support while writing this early version of book.
Signal Reconstruction with AI
Status: Not yet available - Stay tuned!Discovery: A New Approach for Denoising and Reconstruction
Key Insight
Through the combination of attention mechanisms, positional encoding, and CNN-based reconstruction, we have effectively developed a new way to perform denoising and signal reconstruction for time-domain audio signals.
What Makes It Special?
Unlike traditional denoising methods (like spectral subtraction or Wiener filtering), our approach:
- Preserves temporal and harmonic structures (thanks to attention capturing inter-frame relations).
- Uses frame-level processing with overlap-add for smooth stitching.
- Leverages CNNs for local pattern learning and noise suppression.
This hybrid approach combines strengths from deep learning (attention, CNN) with classical signal processing techniques (framing, overlap-add).
Why This is Important
Traditional denoising methods often distort phase or introduce musical noise.
- Our approach keeps the signal-to-noise ratio (SNR) high, while preserving natural tonal quality.
- This is particularly useful in musical note processing (like MiddleC), speech enhancement, or even bio-signal reconstruction.