Sagar Jeevan

Graduate student at Stony Brook University
Graduating in December 2019 | Actively looking for full-time positions

image

I am a Computer Science Master’s student at Stony Brook University. I will be graduating in December 2019. Currently, I am working as a Research Assistant at File systems and Storage Lab under the guidance of Prof. Erez Zadok. I am working on a User Space File System framework called FUSE, where my role is to develop techniques that help improve the performance of the native framework. The courses that I have taken have given me an opportunity to build robust multi-threaded projects such as Stackable File System, Cache Efficient Parallel Scheduler, Loadable System Call by using C, C ++, Python programming languages. These experiences have motivated me to pursue my career on File System, Operating System, Distributed System, Virtualisation, and other related domains.


Work Experiences

Graduate Research Assistant (under Dr. Erez Zadok)

File Systems and Storage Lab | June 2019 - Present

Improving the performance of the Filesystem in Userspace (FUSE) framework by reducing the total trip time taken by requests to get processed. Improvements in the processing of requests are being developed by implementing a zero-copy data transfer to FUSE daemon by creating a shared memory map between a kernel driver queue and FUSE Daemon.

Full Stack Web Developer, learnelements.com​

Nov 2016 - Dec 2017

Initiated, designed, and developed a dynamic website for people to read, write, and share insightful articles.

  • Implemented a user registration and login system coupled with social media authentication.
  • Collaborated discussions by developing RESTful HTTP APIs for threads such as comments, contact me, and feedback.
  • Reinforced article search with text indexing and by implementing hashtags to guide the efficient lookup of articles.
  • Developed proactive security measures for database hacking methods such as SQL Injection, and DoS attacks.
  • Practiced Agile methods such as Iterative and Incremental Development, Continuous Deployment, and Code Refactoring.
Tech Stack: ​Django Web Framework (Python), Bootstrap, jQuery, AJAX, Amazon Web Services​

Freelance Content Writer, unbelievable-facts.com

Jan 2017 - Mar 2017

As a part-timer, researched random facts and published multiple articles and listicles. [The Facebook page has 7.8 million followers]

Research Assistant, RNS Institute of Technology

Jan 2015 - May 2015

Researched a myriad of Image Fusion algorithms during my undergraduate studies to assist a professor in his research on Image Fusion. Experimental results of algorithms were derived and compared to gain insights on the performance metrics of each algorithm.

Projects

Designed and developed a stackable Linux file system that automatically creates a backup file when a regular file is being written, and also allows one to view versions as well as recover and delete. Design for challenges such as when to backup, where to backup, version management, visibility policy, and retention policy were deduced. The file system’s performance gained a maximum of over 70% by making rapid data copies using zero-copy splice compared to the traditional data copy with a user-space buffer.

  • Cache Efficient Parallel Scheduler
  • C++, Cilk Plus, PAPI

    Designed and Developed three cache efficient schedulers for a parallel Matrix Multiplication program. Distributed work-stealing, distributed work-sharing, and centralized work-sharing schedulers were efficiently implemented to optimize performances such as GFLOPS, and L1/L2/L3 cache hits. The overall performance was improved by 350% by a work-stealing scheduler with a centralized scheduler’s results as the benchmark.

  • Linux System Call
  • C, Kernel Programming

    Designed and developed a Linux loadable kernel module that supports a system call for encryption, decryption, and copy of files. A buffer was implemented to efficiently copy, encrypt, and decrypt data using MD5, and AES cipher with CTR mode.

  • Kernel Hacking​
  • C, Kernel Programming

    Developed buggy code with Linux’s kernel hacking configurations enabled, to debug and understand commonly occurred kernel bugs such as stack-overflow, atomicity corruption, locking correctness, memory leak, and memory poison overwritten.

    Designed and developed a stackable Linux file system that automatically creates a backup file when a regular file is being written, and also allows one to view versions as well as recover and delete. Design for challenges such as when to backup, where to backup, version management, visibility policy, and retention policy were deduced. The file system’s performance gained a maximum of over 70% by making rapid data copies using zero-copy splice compared to the traditional data copy with a user-space buffer.

  • Kmer Estimate
  • C++, OpenMP

    Implemented an efficient parallelization of an optimal space usage streaming algorithm that approximates the count of k-mers with a given frequency for large-scale genomic data. The overall performance of the algorithm was improved by 30%.

  • CricBuzz Live Score
  • ​XML, Python

    Contributed to an open-source GitHub repository that fetched live cricket (sport) scores from a cricbuzz.com​ API. Identified existing bugs and developed code for efficient score retrieval.

    Statistical analysis of the Los Angeles payroll data of government employees were performed to gain insights on the distribution of employees’ salary and benefits over years, departments, and career levels. Experimental results were derived by analyzing the data against various testing methods such as Kolmogorov-Smirnov Test, T-Test, Wald’s Test, Permutation Test, and Linear Regression.