Anubhav Jana

WORK HISTORY

EDUCATION

ABOUT ANUBHAV JANA

Currently, I am a Research Engineer at IBM Research Labs India, where I work on production-level llm.build. llm.build is a unified build system for AI model development that ensures reproducibility, traceability, and a standardized process for creating custom AI models. It streamlines workflows and enables collaboration across IBM and IBM Research, helping drive individual innovations into general-purpose open-source LLMs such as Granite.As a Research Fellow at Intel Labs, I focused on AI workload characterization, analyzing and optimizing the performance of large-scale AI applications. My work involved deep-diving into system architectures, identifying bottlenecks, and enhancing efficiency for AI workloads running on modern intel hardware.Previously, as a Hybrid Cloud Intern at IBM Research, I worked on statistical multiplexing of user requests with minimal GPU resources, optimizing cloud efficiency and resource allocation for AI inference workloads.Prior to my internship at IBM Research, as an MS student in Computer Science and Engineering at IIT Bombay, I worked on cloud infrastructure, serverless platform optimization, virtualization, and cloud computing. I had published a paper titled DAGit, a platform for enabling serverless applications, at the International Conference on High Performance Computing (HiPC) 2023, demonstrating my research skills and innovation. This work heavily included Kubernetes and Performance Evaluation and optimisations for Data Flow. Link to my paper : have also gained hands-on experience as an enterprise software engineer and an associate infrastructure engineer at Wolters Kluwer India, where I worked with various cloud technologies and tools, such as AWS, Azure, Terraform, Kubernetes, Linux, PowerShell, Python, Flask, Django, REST API, Git, and Azure Devops. I contributed to the development, deployment, and automation of cloud solutions, as well as the improvement of performance, scalability, and security.