Kureishi Shivanand
Portfolio

Showcase of my projects along with my technical knowledge and abilities.

Projects

This section displays my projects that illustrate my technical skills and knowledge.
The title of the project, a picture correlating to which set of skills it correlates to,
and a description of accomplishments of such project.

Note
There are 5 sets of technical skills that are obvious from the provided picture:
Android Development, Computer Networking, Computer Software, Computer Hardware, Technological Communication.

Engineering Capstone (Android App)

  • Engineered ad-hoc text messaging android app to facilitate multi-device communication
  • Improved coding solutions to be most optimal per client’s requirements
  • Tracked and reported progress via constant reports and meetings to supervisor
  • Organized and led frequent team meetings to discuss direction of project and requirements to be fulfilled

Anonymous Message Broadcaster

  • Established anonymous communication between multiple client and server using multi-threading principles
  • Implemented socket programming using Java Networking and Encryption to simulate network sessions

Routing Control System for Inter-domain Routing

  • Designed controller to compute shortest path to other networks via link costs
  • Generated packet routes through various networks
  • Integrated knowledge of BGP and inter-domain routing

Ticketing System Software

  • Collaborated with team members to create interactive console for ticketing service
  • Executed multiple black and white box testing through vigorous regression testing
  • Integrated Test Driven Development using different levels of testing (unit cases, integration, system etc.)
  • Incorporated Sprint planning within Agile environment
  • Utilized SDLC layout and documented along each stage with improvements

Network File Transfer Application

  • Created client that was able to upload and download file from server using Java Networking library
  • Client request to change directories or list files in directory from server
  • Utilized knowledge in TCP protocols and socket programming

Embedded Systems Media Center

  • Produced an audio player, photo gallery and game on NXP LPC17xx board
  • Enhanced knowledge in programming embedded systems and C using uVision IDE

1-bit Full Adder (IC Design)

  • Designed full-adder circuit using logical effort method to receive a specific load capacitance
  • Developed schematic of circuit and executed parametric and DC analysis to ensure correct functionality
  • Created PCB layout of schematic using Virtuoso Layout Editor Turbo
  • Generated testbench to compare schematic and extracted view from layouts

Cache Controller

  • Programmed cache controller to interface SRAM units with other devices using Xilinx Spartan-3E FPGA
  • Incorporated VHDL coding techniques in Xilinx ISE CAD to implement controller
  • Executed program on FPGA and monitored using built-in performance tools

Semi-RISC CPU

  • Designed and implemented 1-bit semi-RISC CPU on DE2 board
  • Enhanced practical experience using VHDL as an HDL
  • Administered appropriate control signals to data-path elements to achieve desired operation

Function Generator

  • Designed and implemented function generator using Operational Amplifiers
  • Generated desired square/triangle waveform per requirements
  • Incorporated Voltage-controlled frequency, Frequency Range Select and Amplitude Control

eCommerce System

  • Created interactive GUI using Java and JFrame library
  • Designed program flow using UML User and Class Diagrams
  • Implemented design and developed unit test cases for desired functionality
  • Improved object-oriented coding techniques

BJT Amplifier

  • Designed and implemented inverting 50V Amplifier with 20kHz bandwidth using 2 stages
  • Generated functional simulations using NI Multisim
  • Analyzed physical circuit using oscilloscope and multimeter

Coffee Maker Hackathon

  • Redesigned household coffee maker for smart pouring addition
  • Strengthened collaboration and communication skills
  • Improved documentation skills through constant progress reports

Predicting College Admission using R

  • Analyzed historical data for college admission and determined driving factors using R
  • Conducted descriptive analysis on variables in dataset to demonstrate correlation with admission
  • Performed predictive analysis using Logistic Regression, SVM and Decision Trees models on admission
  • Determined significant variables to the target variable as well as choose a champion prediction model

Determine Source of Complaints by Customers using SAS

  • Created trend chart to track number of complaints by day and month
  • Categorized complaint types and determined which occurred the most frequent
  • Visualized percentage of resolved/unresolved complaints by State
  • Investigated if company utilized resources equally to resolve complaints by frequency plot

Building User-based Recommendation Model for Amazon using Python

  • Utilized Python libraries to explore movie rating’s dataset with missing values
  • Extracted basic statistics of dataset to gain deeper insight into its composition
  • Decided on logical method to fill missing values in dataset to allow the training of the model
  • Imported Surprise library to predict user ratings for movies using SVD algorithm
  • Tested the recommendation model and achieved a high prediction accuracy

Mercedes-Benz Greener Manufacturing

  • Utilized Python resources to explore Mercedes-Benz dataset to reduce the time cars spend on the test bench
  • Reduced training features by removing zero-variance columns and applied label encoders to categories
  • Conducted dimensionality reduction on training features using PCA to reduce training time for the model
  • Developed highly efficient and accurate model to predict the time it takes a car to pass the testing

Sales Performance Analysis

  • Utilized Tableau Desktop capabilities to visualize data in comprehensive manner to display to clients
  • Blended separate data sources together to incorporate separate measures into a single visualization
  • Created Bullet Chart using 2 measures from different data sources
  • Implemented calculated field on color to determine if the monthly sales had exceeded the sales target
  • Incorporated filter to show visualizations by the selected year

K-Means Clustering for Telecommunication Domain

  • Utilized Apache Spark to implement K-Means cluster algorithm on telecommunication complaint dataset
  • Incorporated Scala language in order to interact with Apache Spark through spark-shell
  • Evaluated clusters based on training data and tweaked input parameters to improve model accuracy

Data Science Capstone (Diagnosing Diabetes in Patients)

  • Pre-processed data by imputing missing data with variable mean
  • Performed descriptive analysis by visualizing utilizing histograms, scatter plots, bubble charts and pie charts
  • Conducted correlation analysis using a heatmap to determine factors that affect diabetes diagnosis
  • Devised strategies to build a prediction model to classify whether a patient had diabetes
  • Decided on cross-validation as an appropriate validation framework
  • Tested various models using recall, AUC(ROC) curve, accuracy score etc. to choose the most accurate model

Lending Club Loan Data Analysis

  • Utilized TensorFlow and Keras to implement ANN to classify if a customer meets loan criteria
  • Conducted EDA on all variables to analyze their distribution and assess if outliers were present
  • Performed Feature Transformation and Extraction on dataset to allow faster computation of the neural network
  • Optimized hyperparameters using GridSearchCV to achieve an acceptable classification accuracy

Artificial Intelligence Capstone (Finance – Credit Card Fraud)

  • Performed Exploratory Data Analysis (EDA) on entire dataset
  • Utilized resampling techniques to remedy imbalanced dataset in during model training
  • Assessed Naïve Bayes, Logistic Regression and SVM models using various evaluation metrics
  • Tuned in-built parameters of tree-based models to help with the imbalanced dataset
  • Conducted hyper-parameter tuning on an ANN model to attain the highest performance
  • Implemented anomaly detection systems to differentiate fraud transactions from valid ones
  • Transformed anomaly scores to engineered features to re-assess models for difference in performance