TA

Tanuj Aasawat

AI Research Scientist

Tanuj Aasawat - AI Research Scientist, Intel Labs



Profile Links:




Current Position:


AI Research Scientist at Intel Labs, Bengaluru, India

Professional Background:


Education:
  • Master of Applied Science (MASc), Computer Science (2015-2018): University of British Columbia

  • Advisor: Professor Matei Ripeanu

  • Thesis: Scale-Free Graph Processing on a NUMA Machine

  • Achievements: Developed a NUMA-architecture aware graph processing framework. Secured World Rank 2 in Graph500 challenge (June 2018).


Previous Experience:
  • Researcher (Oct 2018 - Present): RIKEN Center for Advanced Intelligence Projects, Tokyo, Japan.

  • Focus: High Performance Computing (HPC) and Applied AI

  • Achievements: Developed a novel massively parallel Monte Carlo Tree Search (MCTS) for de novo Molecule Generation, HyGN: Hybrid Graph processing engine for NUMA, among others.


  • Graduate Research Assistant (Sep 2015 - Sep 2018): University of British Columbia

  • Key Contributions: Designed and implemented a high-performance NUMA-aware graph processing framework. Multiple top performances in Graph500 challenges.


  • Internships:

  • Amazon Web Services, Vancouver (Fall 2016): Designed and implemented a system for AWS RDS-Oracle clients.

  • IBM Research - Almaden, San Jose (Summer 2016): GPU acceleration for Deep Learning in Apache SystemML.

  • Mitacs Globalink Research Intern, University of British Columbia (Summer 2014): Parallel clustering coefficient computation algorithm on GPUs.

  • 3v Geomatics Inc, Vancouver (June - July 2014): Accelerated a linear algebra kernel in InSAR processing pipeline on GPU.

  • Indian Institute of Technology, Kharagpur (May - June 2013): Developed models for language evolution using Genetic Algorithms.


Research Interests and Notable Projects:


  • High Performance Computing (HPC)

  • Large-scale Graph Analytics

  • Applied Artificial Intelligence


Significant Research Contributions:
  • Monte Carlo Tree Search (MCTS) applied to real-world problems and molecule generation.

  • NUMA-Architecture Graph Processing Framework: Achieved exceptional performance gains, recognized in global Graph500 rankings.

  • Publications:

  • Scale-Free Graph Processing on a NUMA machine: ACM/IEEE Workshop on Irregular Applications: Architectures and Algorithms (2018).

  • Performance Analysis on CPU, GPU, and Hybrid Graph Processing Frameworks: IEEE International Workshop on High-Performance Big Data (2018).

  • Efficient GPU Techniques for Processing Satellite Image Data: Super Computing Conference (2015).

  • Other technical papers and posters presented in reputable conferences.


Awards and Recognitions:


  • Hult Prize Regional Finalist (2018)

  • UBC Graduate Research Assistantship (2015 - Present)

  • ACM/IEEE Travel Grants for SuperComputing Conferences (2014, 2015, 2017)

  • International HPC Summer School Participant, Boulder, Colorado (2017): Selected from prestigious international university representatives.

  • Mitacs Globalink Graduate Fellowship (2015)

  • Multiple Merit-Cum-Means Scholarships during academic tenure


Professional Skills and Technologies:


  • Programming: C/C++, Java, CUDA, OpenMP, PThreads

  • Technologies: AWS, Cloud Computing, Nvidia GPUs, JCuda, cuBLAS, cuDNN

  • Specialized Competencies: NUMA architecture, large-scale graph analytics, high-performance computing frameworks


Publications and Presentations:


Tanuj has published several impactful papers focusing on advanced graph processing and high-performance computing in renowned journals and conferences. This includes contributions to the ACM/IEEE workshops and other notable scientific gatherings.

Analysis:


Tanuj Aasawat brings considerable expertise in HPC and AI, making him a valuable contact within Intel Labs. His background in developing cutting-edge graph processing frameworks and deep learning acceleration demonstrates his ability to contribute to high-stakes, innovative projects. Leveraging his experience and publications could offer strategic insights and potential collaborations in advanced computational research and development initiatives.