The Computing Systems research team at TCS focuses on the development of cost effective, high performing fault tolerant systems for modern applications. Apart from analytics and ML/DL, it also includes engineering research leading to the development of performance models, middleware and frameworks for accelerating development and deployment of data driven and transactional applications. The team also works on the benchmarking and analysis of bleeding-edge hardware and software technologies to identify applications that can benefit from such frameworks.
The sub-areas active within this research area include:
- Performance for AI: This team is focused on building scalable high-performance architectures for ML/DL/RL pipelines, including both training and inference, by accelerating data management, innovations in ML/DL/RL algorithms, high-performant feature stores, and hybrid deployments using serverless architectures.
- AI for Performance: In this group, they focus on use of ML/DL/RL algorithms to model and accelerate system performance
- Special-purpose hardware and FPGA: This team keeps track of the latest hardware developments and design custom hardware to enhance the performance of enterprise systems
- Performance modeling and optimization: This group works on building highly accurate predictive performance models by considering a number of applications, systems, and workload characteristics, and optimizing workload and system configurations to enhance the overall system performance by leveraging these models
- Quantum computing: This team works on identifying algorithms and applications that are likely to benefit from quantum computing in the near future, in addition to design systems interfacing with quantum computers
People & Patents
Academic partners: Tata Institute of Fundamental Research, India
Patents and publications (2016 onwards): This team has been granted more than 10 patents and has had over 40 papers published at key conferences and journals