Portrait of Divya Mahajan

Divya Mahajan

Senior Researcher

About

I am a Senior Researcher at Microsoft Azure’s Cloud Accelerated Systems & Technologies team since September 2019. My research interests lie in devising next-generation sustainable compute platforms targeting end-to-end data pipeline for large scale AI and machine learning. My work draws insights from a broad set of disciplines such as, computer architecture, systems, and databases.

I received my Ph.D. from Georgia Institute of Technology in 2019 and obtained a Master’s  from The University of Texas at Austin (2014). I obtained my Bachelor’s from Indian Institute of Technology, Ropar where I was conferred the Presidents of India Gold Medal, the highest academic honor in IITs.  My research has been published in top-tier venues such as ISCA, HPCA, MICRO, ASPLOS, NeuRIPS, and VLDB. My dissertation work  has been recognized with the NCWIT Collegiate Award, 2017 and distinguished paper award at High Performance Computer Architecture (HPCA), 2016. I continue to serve in the Program Committee for ISCA, HPCA, MICRO, MLSys, IISWC, and FPGA.

  • Integrating Accelerators within Database Management Systems for Advanced Analytics
    Devised a cohesive system by understanding the horizontal pipeline of systematic data processing. We developed a vertical full-stack solution by addressing the challenges observed in executing advanced analytics within an RDBMS setting.
  • Template-based Framework for Accelerating Machine Learning
    We devised a comprehensive solution, from programming model down to circuits, to automatically generate accelerators for machine learning including deep learning.
    Tabla: http://act-lab.org/artifacts/tabla/
    DnnWeaver: http://dnnweaver.org/
    PolyMath: https://github.com/he-actlab/polymath
  • Systematic architectural and compilation techniques to control quality tradeoffs in Approximate Computing
    Hardware-software techniques with components in both compiler and microarchitecture to transforms the tradeoff between quality and energy-performance gains from approximation.

My research interests lie at the intersection of computer architecture and systems for large-scale data intensive applications.