About Me
My name is Kesavar Kabilar. I’m an efficiency-driven data science enthusiast with expertise in databases and machine learning. I enjoy coding solutions that optimize business processes and leverage data to drive actionable insights.
I am working at Oracle as a Software Engineer (Data), optimizing queries, scripts, and web services. I'm also currently pursuing a Master of Computer Science at the University of Illinois Urbana-Champaign, specializing in Data Science.
In my free time, I like to play chess (2300 Rapid rating), solve a Rubik's Cube (average around 14 seconds), and juggle (5 balls for 20 seconds).
Work Experience
Oracle NetSuite
Software Engineer - Data Integration and Data Automation
Apr 2024 - Present
Skills: SQL, JavaScript (SuiteScript), Web Services (REST and SOAP)Â
Develop 30+ SQL queries to optimize data extraction processes using the Oracle Data Integrator API.
Streamlining data retrieval from Oracle’s database, and accelerating report generation by approximately 13% for critical business insights.
Automated key business processes by developing 50+ SuiteScripts (JavaScript), including scheduled and map-reduce scripts. These automated tasks involved invoice generation, receipt creation, and billing record updates, resulting in approximately 8% efficiency improvement in NetSuite operations.
Enhanced data integration capabilities by leveraging 20+ NetSuite's REST and SOAP API web services to integrate with 10+ external systems (Power BI/Boomi/Snowflake). This expanded system interoperability and facilitated seamless data exchange.
Community Legal Education Ontario
Software Engineer - Quality Assurance
Sep 2023 - Mar 2024
Skills: Python, Selenium, Unit Tests, Scrum
Reduced post-release issues by 15% through extensive testing of CLEO's legal documentation software, using Python and Selenium for quality assurance.
Boosted testing efficiency by 10%, streamlining bug identification during development with industry-standard tools.
Lowered critical defects in final releases by 22% across 10 projects, cutting development-to-production time by 14% through effective test plans and collaboration with developers.
ICUBE UTM
Software Engineer Intern
May 2020 - Aug 2022
Skills: PyTorch, Numpy, RNN, MongoDB, Express.js, React.js, Node.js, Jira
Developed an RNN model in PyTorch for stock price prediction, achieving 0.071% prediction accuracy after optimizing hyperparameters like Epochs, Batch Size, and Previous Days through 50+ tests.
Led MERN stack implementation to optimize client engagement with Bitcoin, Blockchain, and NFT management, increasing user satisfaction by 20%, reducing backend response time by 15%, boosting client engagement by 25%, and extending user session duration by 30%, while achieving project milestones 10% ahead of schedule.
Education
University of IllinoisÂ
Urbana-Champaign
Master of Computer Science - Data Science
Aug 2024 - May 2026
Coursework: Data Mining, Data Visualization, Applied Machine Learning, Natural Language Processing, Deep Learning for Healthcare, Cloud Computing Application, Scientific Visualization, and Theory and Practice of Data Cleaning.
University of Toronto
Honors Bachelor of Computer ScienceÂ
Sep 2019 - Jun 2023
GPA: 3.87/4.00 (High Distinction), Dean's List of Scholar 2020-2023
Coursework: Machine Learning, Neural Network and Deep Learning, Computer Vision, Algorithm Design and Analysis, SQL Databases, Data Structures and Analysis, Linear Algebra, Probability and Inductive Logic, Numerical Methods
Projects
Apache HBase, Apache Zookeeper, And Apache Phoenix
Set up and configured a data management system using HBase, ZooKeeper, and Phoenix within Docker, focusing on storing, managing, and querying data.
Created HBase tables, loading data, implementing Java-based SQL join simulations, and utilizing Phoenix for SQL queries against the HBase database.
AWS RDS Aurora MySQL and AWS ElastiCache
Built a serverless storage service on AWS, using Lambda, Aurora MySQL, and ElastiCache (Redis), to demonstrate how caching improves database performance.
Implemented caching strategies (write-through and lazy-loading) within a Lambda function that manages read and write operations between an Aurora MySQL database and a Redis cache.
PySpark MapReduce
This project explores the analysis of Wikipedia datasets, including page titles and link structures, using big data technologies.
Docker containers provide a standardized development environment for consistent execution.
Apache Spark, a distributed computing system, is used to implement MapReduce algorithms for efficient data processing.
AWS Lex V2 & Lambda
Implemented a chatbot using AWS Lex that calculates the shortest distance between two cities in a directed graph.Â
The project utilizes various AWS services, including: AWS Lambda, AWS API Gateway, AWS DynamoDB, AWS Lex, AWS Cognito
AWS Load Balancing + Auto Scaling
Two Amazon EC2 instances were provisioned to distribute the application workload.Â
An AWS Auto Scaling group was implemented to dynamically manage capacity, ensuring application availability and preventing overload
Apriori Algorithm for Frequent Itemset Mining
Implemented the Apriori algorithm to discover frequently occurring category combinations from a real-world dataset.
Contiguous Sequential Pattern Mining
Discovered frequently occurring, consecutive sequences of words (phrases) within a dataset of text reviews.
Hierarchical Clustering with Agglomerative Approach
Applied an agglomerative hierarchical clustering algorithm to group geographical data points based on their similarity.
Clustering Validation
Evaluated cluster results by utilizing Jaccard Similarity and Normalized Mutual Information (NMI).Â
Decision Tree
Built a decision tree by calculating information gain on each attribute to find the most influential attribute at each level.Â
Naive Bayes Classifier
Computed naive nayes classifier to predict the animal class (mammal, bird, reptile, etc.) based on various attributes like hair, feathers, eggs, etc., using the zoo animal classification dataset.Â
Developed a Recurrent Neural Network (RNN) model using PyTorch, NumPy, and Matplotlib for stock price prediction based on historical data. The model achieved an accuracy rate with minimal deviation of 0.071% in stock price predictions.
Achieved maximum accuracy in stock price prediction through various optimization processes involving 50+ tests on numerous hyperparameters such as Epochs, Batch Size, and Number of Previous Days.
Accomplished a 20% increase in user satisfaction by leading the MERN stack implementation for optimizing client engagement in learning, tracking, and managing Bitcoins, Blockchains, and NFTs.
Enhanced backend efficiency with a 15% reduction in response time for "GET," "POST," and "PUT" operations by developing and integrating Restful APIs. This optimization significantly maximized the overall performance of backend processes.
Engineered visually stunning web pages, resulting in a 25% boost in client engagement and a 30% extension in average user session duration, showcasing outstanding teamwork with a 10% ahead-of-schedule achievement of project milestones.
Certifications
Proficient in virtualization, containerization (specifically Docker), Dockerfile creation, multi-container orchestration with Compose and Airflow, Kubernetes core concepts, cluster architecture, deployment using cloud environments, GitHub Codespaces, and AI-driven tools, and effectively handle data scenarios through mastering containerization, deploying apps, and addressing production issues with cloud orchestration and SRE practices.