With nearly a decade and a half of experience in the IT industry, I have been primarily working in the Application development and maintenance area of almost all major retailers in the USA. Although the experience of working in the retailing industry has been rewarding & fulfilling but off late, I was becoming more inquisitive towards new emerging technologies and the rapid adaptation of digital technologies by almost all companies in all the industries. Artificial intelligence and Machine learning were at the forefront of the emerging digital technologies, and I got reskilled in this area with Great Learning in the PGP program for AI & ML.
Armed with the knowledge of this cutting-edge technology, I got aligned to the emerging Artificial Intelligence unit within my company, where I was tasked to drive a major R&D effort for a major telecom company in the USA on their efforts in providing Accessibility services to the Visual and Hearing-impaired customers. I had to develop a prototype for identifying everyday household sounds in the customer’s home like smoke detector alert, doorbell ringing, microwave beeping, water running, dog barking, cat meowing, conversations, footsteps, etc.
I implemented a custom deep learning model based using Google Yamnet transfer learning. The model had to be lightweight as it had to be implemented in the devices at the customer’s home. I performed model optimizations after multiple rounds of iteration to develop a model with high test accuracy for all the different sounds and optimized for memory. The model was then implemented in an edge inferencing device which will continuously record the incoming sounds and alert on various sounds to the customers.
The whole experience from problem definition to implementing the model to the edge device has been deeply satisfying as it would help the hearing-impaired customers to better understand the various sounds happening in their own home and to which they have been completely oblivious to had it not been to the recent advance in AI & ML.
The field of AI/ML is very vast, and one needs to constantly be aware of the latest innovations in this field. I learn new concepts, frameworks, techniques & tools to stay abreast of the recent advancements in this field. While this constant endeavor to keep oneself updated is challenging, I can attest that the rewards are even greater both personally and professionally.
