arun k rajagopalan - parasol laboratory · arun k rajagopalan, jeff huang acceptance rate: 27%...

1
Arun K Rajagopalan linkedin.com/in/arunxls; github.com/arunxls 401 Stasney Street, Apt 511, College Station, TX 77840 [email protected] (979)-985-0859 Texas A&M University, College Station, Texas Master of Science in Computer Science GPA: 3.80 / 4.0 Advanced Compilers Design, Parallel Algorithms, Software Engineering, Analysis of Algorithms, Reliable Concurrent Software 2016 TECHNICAL SKILLS Programming Languages:: Java, Perl, C, C++, Python, Ruby, JavaScript Tools:: Ruby on Rails, Shell Scripting, Matlab, LLVM Competitions:: In the 90 th percentile on HackerRank under Algorithms - hackerrank.com/arunxls EDUCATION National Institute of Technology (NIT), Tiruchirapalli, India Bachelor of Technology in Electronics and Communications Graduated First Class GPA: 7.93 / 10 ARM System Architecture, Image Processing, Networks and Protocols, Advanced Microprocessors 2012 EXPERIENCE Software Engineering Intern, Nvidia Graphics Private Ltd. Westford, Massachusetts Revamped system design for a job resource predictor used to schedule jobs on the compute farm. MySQL schema redesign using of indices, stored procedures and normalized fields. Gathering and generating metrics to track health of the system and provide feedback to users. 2015 Architect, Infrastructure and Tools, Nvidia Graphics Private Ltd. Bangalore, India Developer on an application that enables users to write, manage and execute regression testing. Ensured efficient and fair utilization of the distributed compute farm cluster to all users. Various improvements to the parser that improved launch time, maintenance and scalability. 2012 - 2014 PUBLICATION RDIT: Race Detection from Incomplete Traces Arun K Rajagopalan, Jeff Huang Acceptance rate: 27% Designed and implemented an enhancement to the Happens-Before algorithm that guarantees preciseness in situations when sections of the source code are unavailable. We observed that our enhanced algorithm completely eliminated false positives when run against Apache’s Derby DB. ESEC/FSE NIER 2015 PROJECTS Source to source optimizing Fortran-77 compiler Optimizing pass that automatically parallelizes a program using OpenMP pragmas. Data dependence analysis was performed on the SSA form to ensure no dependency violations. Parallel dynamic race detection Enhanced program instrumentation using thread-safe data-structures and a master DB. Improved analysis time by diving the workload among workers in a thread-pool. 2015 Evaluation of parallel algorithms on OpenMP and MPI Implemented two versions of the parallel-scan on shared memory and message passing interface. Studied various parameters such as weak/strong scaling, performance and complexity. 2015 Dynamic Instrumentation Framework for C++ Built an LLVM pass that intercepted read/write, lock/unlock and fork/join calls into a log file. The generated log file was then fed to standard race detection tools. 2014 Construction of Maximum Bandwidth Paths Maximum bandwidth path construction was implemented using Kruskal’s and Djikstra’s algorithm. Kruskal’s algorithm was found to have better performance as graph density increased. 2014 AggieExperts: A content management system for Texas A&M faculty Created the public-visible profile view and user-visible profile edit and creation pages. Client-side data validation using Javascript and server side data retention using PostgreSQL. Test-driven development using RSpec and Cucumber. 2014 TAMUHack: WebBro Winner of Microsoft’s ‘best’ hack award for innovative use of the Azure platform. WebBro allowed users to add comments to websites to leave feedback to owners and visitors. 2014 2015

Upload: hoangthien

Post on 20-Jan-2019

224 views

Category:

Documents


0 download

TRANSCRIPT

Page 1: Arun K Rajagopalan - Parasol Laboratory · Arun K Rajagopalan, Jeff Huang Acceptance rate: 27% Designed and implemented an enhancement to the Happens-Before algorithm that guarantees

Arun K Rajagopalan linkedin.com/in/arunxls; github.com/arunxls401 Stasney Street, Apt 511, College Station, TX 77840

[email protected](979)-985-0859

Texas A&M University, College Station, TexasMaster of Science in Computer Science GPA: 3.80 / 4.0Advanced Compilers Design, Parallel Algorithms, Software Engineering, Analysis of Algorithms, Reliable Concurrent Software

2016

TECHNICAL SKILLS

Programming Languages:: Java, Perl, C, C++, Python, Ruby, JavaScriptTools:: Ruby on Rails, Shell Scripting, Matlab, LLVMCompetitions:: In the 90th percentile on HackerRank under Algorithms - hackerrank.com/arunxls

EDUCATIONNational Institute of Technology (NIT), Tiruchirapalli, IndiaBachelor of Technology in Electronics and Communications Graduated First Class GPA: 7.93 / 10ARM System Architecture, Image Processing, Networks and Protocols, Advanced Microprocessors

2012

EXPERIENCE

Software Engineering Intern, Nvidia Graphics Private Ltd.Westford, Massachusetts

Revamped system design for a job resource predictor used to schedule jobs on the compute farm.MySQL schema redesign using of indices, stored procedures and normalized fields.Gathering and generating metrics to track health of the system and provide feedback to users.

2015

Architect, Infrastructure and Tools, Nvidia Graphics Private Ltd.Bangalore, India

Developer on an application that enables users to write, manage and execute regression testing.Ensured efficient and fair utilization of the distributed compute farm cluster to all users.Various improvements to the parser that improved launch time, maintenance and scalability.

2012 - 2014

PUBLICATION

RDIT: Race Detection from Incomplete TracesArun K Rajagopalan, Jeff Huang Acceptance rate: 27%Designed and implemented an enhancement to the Happens-Before algorithm that guarantees preciseness in situations when sections of the source code are unavailable. We observed that our enhanced algorithm completely eliminated false positives when run against Apache’s Derby DB.

ESEC/FSE NIER 2015

PROJECTS

Source to source optimizing Fortran-77 compilerOptimizing pass that automatically parallelizes a program using OpenMP pragmas.Data dependence analysis was performed on the SSA form to ensure no dependency violations.

Parallel dynamic race detectionEnhanced program instrumentation using thread-safe data-structures and a master DB.Improved analysis time by diving the workload among workers in a thread-pool.

2015

Evaluation of parallel algorithms on OpenMP and MPIImplemented two versions of the parallel-scan on shared memory and message passing interface.Studied various parameters such as weak/strong scaling, performance and complexity.

2015

Dynamic Instrumentation Framework for C++Built an LLVM pass that intercepted read/write, lock/unlock and fork/join calls into a log file.The generated log file was then fed to standard race detection tools.

2014

Construction of Maximum Bandwidth PathsMaximum bandwidth path construction was implemented using Kruskal’s and Djikstra’s algorithm.Kruskal’s algorithm was found to have better performance as graph density increased.

2014

AggieExperts: A content management system for Texas A&M facultyCreated the public-visible profile view and user-visible profile edit and creation pages.Client-side data validation using Javascript and server side data retention using PostgreSQL.Test-driven development using RSpec and Cucumber.

2014

TAMUHack: WebBroWinner of Microsoft’s ‘best’ hack award for innovative use of the Azure platform.WebBro allowed users to add comments to websites to leave feedback to owners and visitors.

2014

2015