Samudyatha Kaira

Experience:

• Developed AI models for tabular data using deep learning and classical machine learning methods (pytorch, sklearn). Delivered supervised learning model of >90% accuracy to large industrial customers.
• Designed and developed Edge to cloud architecture to scale AI Edge solution across multiple edge devices in partnership with large scale cloud customer(s) using cloud native tools such as Kubernetes (Enterprise), docker, git, Cloudbuild, helm, etc.
• Designed and developed AI edge solutions for industrial customers from pilot to production. Owned technical architecture of industrial edge solutions platform for time-series data management and inference stack.
• Developed performance benchmarking end to end AI workloads on CPU and GPU (Intel and NVIDIA hardware)
• Mentored Machine learning engineers to ramp up on tools to handle large datasets, feature selection, extraction and EDA for successful model development.
• Collaborated with cross-functional teams that include account team to incorporate customer requirements, to silicon team to deliver requirements to silicon roadmap.

Thought leadership

• Article: Co-authored IEEE Internet of Things Magazine Magazine Article https://ieeexplore.ieee.org/author/37089378094
• Patents: Filed 4 US Patents in the AI domain https://patents.google.com/?q=(kaira)&inventor=samudyatha&oq=samudyatha+kaira
• Conference: Led demo of Edge to Cloud in collaboration with Google at Hannover Messe 2022 in Hannover, Germany
• Conference: Delivered talk at Intel Innovation ON 2022 on "Managing data at edge"

• Developed credential management system of then Intel "Secure Device OnBoard" (SDO) now called FIDO (https://fidoalliance.org/device-onboarding-overview). Enabled numerous SDO demos for internal and external customers, including Intel Developer Forum
• Role modeled software engineering best practices: contributed to release planning and led sprint planning, build system setup, code coverage, unit tests.