Yugarsi Ghosh Email & Phone Number

Kirkland, Wa, Us

Seattle, Washington, United States

Greater Seattle Area

Amazon.com, Data Normalization Systems (DNS)• Integrated a Machine Learning based Key-phrase Extraction API into DNS for improving Normalization Rate.• Developed Unit Conversion feature into DNS to improve price per unit (PPU) experience in Amazon.com• Optimized a service to extract production logs and compute deltas between Gamma and Production stages for submissions to the Data Normalization System (DNS). It reduced the running time from 14 hours to 1 hour.• Designed and… Show more Amazon.com, Data Normalization Systems (DNS)• Integrated a Machine Learning based Key-phrase Extraction API into DNS for improving Normalization Rate.• Developed Unit Conversion feature into DNS to improve price per unit (PPU) experience in Amazon.com• Optimized a service to extract production logs and compute deltas between Gamma and Production stages for submissions to the Data Normalization System (DNS). It reduced the running time from 14 hours to 1 hour.• Designed and implemented a 2 layer caching architecture consisting of (L1 level Guava Loading cache and L2 Guava Local cache) to reduce latencies of the rules store of DNS by 30%.• Optimized the conflict checking architecture of mapping uploads in DNS and reduced the time consumption by 5x.• Developed a GUI for configuring rules in the Normalization Rule management services to be used by Operations support using a React-JS based Front End and A-REST backend. Show less

Tempe

Greater Seattle Area

Arizona State University Tempe Az

The Formative Assessment with Computational Technologies (FACT) Lab aims to make improvements in the field of computer-assisted learning and teaching. It deals with design and development of the analytical engine—and the productive integration of technological tools into classroom activities. As a Graduate Research Assistant, I work as a software developer on the Java based Server application and the Android based client application.

Bengaluru Area, India

• Designed Load/Stress test framework for Multi-threaded E-commerce Feed upload into Unbxd Servers using Python.• Developed REST- APIs for generating revenue details from UNBXD API hits using Java.• Collaborated with a team of three people to design a site monitor tool for analyzing the performance of Unbxd Servers using Java.

Bangalore

• Part of Citrix CloudBridge Team.• Develop and maintain the product’s Automated Build Validation Test framework for validation of everysoftware builds released each day.• Designed a deployment framework for Auto-Configuring any complex deployments consisting of severalappliances e.g. routers, switches, clients, servers, and CloudBridge devices in Citrix Network Lab.Software Development Engineer 2CITRIX SYSTEMS R&D INDIA (CITRIX Cloud Bridge(CB) Automation Team) July… Show more • Part of Citrix CloudBridge Team.• Develop and maintain the product’s Automated Build Validation Test framework for validation of everysoftware builds released each day.• Designed a deployment framework for Auto-Configuring any complex deployments consisting of severalappliances e.g. routers, switches, clients, servers, and CloudBridge devices in Citrix Network Lab.Software Development Engineer 2CITRIX SYSTEMS R&D INDIA (CITRIX Cloud Bridge(CB) Automation Team) July 2012 – February 2015• Developed tools for optimizing the memory consumption of various processes in Cloud Bridge(CB) toreduce memory leaks using Python and C++.• Devised a tool for generating fragmented TCP/UDP fragmented traffic packets, to stress test CBperformance on all system ports using Python. Show less

Bengaluru Area, India

• Developed tools for optimizing memory consumption of various process in CloudBridge(CB) to reducememory leaks using Python and C++.• Devised a tool for generating fragmented TCP/UDP fragmented traffic packets, to stress test CBperformance on all System ports using Python.

Kolkata Area, India

Wavelet transform, is a very powerful technique, which helps in multi-resolution analysis of data. Large amount of Rainfall, Temperature, Tree Ring width data of Perth region (West Australia) were collected and de-noised using Discrete Wavelet (Db4). Then the data were subject to regression analysis using MATLAB. The data was further analyzed using multi-resolution, continuous Morlet Wavelet for extracting periodic variations. Periodicities of 78 years were significantly present in almost… Show more Wavelet transform, is a very powerful technique, which helps in multi-resolution analysis of data. Large amount of Rainfall, Temperature, Tree Ring width data of Perth region (West Australia) were collected and de-noised using Discrete Wavelet (Db4). Then the data were subject to regression analysis using MATLAB. The data was further analyzed using multi-resolution, continuous Morlet Wavelet for extracting periodic variations. Periodicities of 78 years were significantly present in almost all sets of data. It was inferred that global warming and cooling in the last century predominantly coincides with a 78 years cycle. • Published a Research Paper on Wavelet Transform in the proceedings of National IEEE conference Calcutta 2011(CALCON 2011).• Published a Chapter in a Book on Data Characterization with Wavelet Transform using MATLAB by CRC Publications. Show less

Roorkee

Designed and implemented a codebook which learns from numerous document files of different types from the web to characterize the average symbol probabilities of various ASCII characters, to build an adaptive Huffman dictionary for data compression. Studied its efficiency and compared its performance with LZW encoding after implementing the same using C++. Also worked on Vector quantization to design a codebook for optimal image compression using MATLAB. Various images were divided into… Show more Designed and implemented a codebook which learns from numerous document files of different types from the web to characterize the average symbol probabilities of various ASCII characters, to build an adaptive Huffman dictionary for data compression. Studied its efficiency and compared its performance with LZW encoding after implementing the same using C++. Also worked on Vector quantization to design a codebook for optimal image compression using MATLAB. Various images were divided into several training vectors. The codebook characterizes the centroid value for all vectors, using clustering algorithms by minimization of mean square error (MSE) between codebook vectors and training vectors after learning from several images adaptively. Show less

Master’S Degree, Computer Engineering, 3.83/4

Activities and Societies: 1.Graduate Student Researcher at CySiS (Cyber Socio- Intelligent ) Lab. 2. Graduate Researcher at FACT lab, ASU..

Bachelor’S Degree, Electronics And Communications Engineering

Education record