software engineer

Pritam Suwal  • Peerlist

Worked on two high-performance computing projects focusing on simulating fish school behaviour using parallel computing techniques. The first project involved implementing both sequential and parallel versions of a fish school simulation programme on the Setonix platform, analysing the impact of various parameters, including thread scheduling methods, and evaluating performance metrics such as speedup and efficiency. The second project utilised MPI and OpenMP to optimise fish data distribution and parallelize simulation steps across multiple processes, addressing challenges like data integrity, parallelization overhead, and bottlenecks. These projects enhanced my understanding of parallel computing, data communication, and performance optimisation in distributed systems.

Optimising Parallel Fish Simulation with MPI 

In this project, I worked on cleaning and profiling datasets ranging from global economic indicators to Olympic medal counts. I designed and implemented a star schema and used ETL processes to load the data into an OLTP system, allowing for efficient querying and analysis. This project provided valuable insights for clients like the National Olympic Committee and the Australian Government, exploring how economic factors and national health metrics influence Olympic performance.

Data Warehouse and ETL Process for Olympic an

An interactive web-based "Sunburst Chart" to efficiently manage a family tree, facilitating connections and communication among dispersed family members. The platform will empower administrators to dynamically add, modify, and update elements on the Sunburst Chart. Additionally, website visitors can view the chart, interact with individual components, and submit comments related to specific chart elements.

UWA Sunburst Family Tree

As part of my coursework in GENG5505, my team and I conducted a comprehensive analysis of the Kidston Solar Project Phase 1 (KS1) as a case study in the project management lifecycle. This project, managed by Genex Power, was a pioneering initiative in North Queensland, aiming to generate 50MW of clean solar energy annually. Our analysis covered all stages of the project management lifecycle—conceptualisation, planning, execution, and finalisation—highlighting key strategies, challenges, and outcomes. We examined aspects such as scope management, risk mitigation, stakeholder engagement, and quality control. Through this study, we gained valuable insights into the complexities and best practices of managing large-scale renewable energy projects. The success of KS1 in terms of both economic impact and environmental sustainability, along with the lessons learned from the challenges faced, provided a rich learning experience that I am excited to apply in future roles within the renewable energy sector.

Kidston Solar Project Phase 1 (KS1) - Project

Collaborated on the development of a Pictionary-inspired web application designed for friends to draw, guess, and have fun together. The project involved both frontend and backend development using HTML, CSS, JavaScript (AJAX, jQuery), and Bootstrap on the client side, with Flask and SQLite on the server side. Successfully set up a test environment, implemented unit and system tests, and utilised Flask's migration tools for database management. The application is hosted on Render.com, showcasing my skills in full-stack development and cloud deployment.

Pictionary: Agile Web Development Group Proje

In this project, I applied machine learning techniques to two distinct datasets: the Breast Cancer Wisconsin dataset and an airline safety dataset from FiveThirtyEight. The goal was to enhance predictive accuracy and derive meaningful insights through data analysis and model building. For the Breast Cancer Diagnosis task, I developed logistic regression and decision tree models to predict whether a tumour is malignant or benign, based on cell nucleus features. The analysis included feature importance determination, PCA-based dimensionality reduction, and a comprehensive comparison between the two models to recommend the most suitable for clinical applications. The Airline Safety Analysis involved clustering techniques on past airline safety records. I implemented K-means clustering to group airlines based on safety metrics, utilising silhouette scores to determine the optimal number of clusters. The project included data scaling considerations, variable selection, and a detailed comparison of cluster characteristics across different time periods.

Machine Learning Project: Breast Cancer Diagn

This report presents an in-depth analysis and implementation of machine learning models using various datasets to explore classification and regression techniques. It is divided into three key sections: • Breast Cancer Wisconsin (Diagnostic) Dataset: The report begins with the exploration of the Breast Cancer Wisconsin dataset, where multiple decision tree classifiers are trained and compared. The tasks include visualising data to identify and drop redundant features, splitting the dataset into training and testing sets, and experimenting with different hyperparameters such as `max_depth`, `min_samples_split`, and `min_samples_leaf`. A grid search with 3-fold cross-validation is conducted to identify the optimal hyperparameters, and the final model's performance is evaluated using confusion matrices, precision, and recall metrics. • Concrete Slump Test Dataset: The second section focuses on the Concrete Slump Test dataset, which is used to build regression models. The report details the data cleaning process, the application of three base estimators (Linear SVM, Linear Regression, and Stochastic Gradient Descent), and the training of a Voting Regressor. The performance of these models is evaluated using RMSE, and the effect of hyperparameter tuning on model accuracy is discussed. • Abalone Dataset: The final section examines the Abalone dataset, utilising "Decision Tree Regression" and "Random Forest Regression" to predict the age of abalones. The report includes steps such as data preprocessing, hyperparameter tuning using cross-validation, and feature importance analysis. The performance of the models is compared, and a Bagging Regressor with SVM regressors as base estimators is implemented to further enhance prediction accuracy.

Optimizing Machine Learning Models: A Compreh

The project report focuses on various cybersecurity challenges and topics, covering a broad range of areas including computer architecture and networking, cryptography, forensics, and vulnerabilities. The report is structured into multiple sections, each dealing with different aspects of cybersecurity, using a fictional narrative centred around "Penguin OS" as a thematic element. 1. Computer Architecture and Networking: This section discusses network security, with a specific focus on finding and exploiting vulnerabilities in network services such as FTP. It involves tasks like using tools (e.g., Nmap) to discover open ports and services, performing anonymous logins to access sensitive information, and SSHing into servers using obtained credentials. 2. Cryptography: The cryptography section presents various challenges that involve decrypting encoded messages using different cryptographic techniques, such as PGP decryption and RSA decryption. Each task is designed to teach different cryptographic principles, such as encryption algorithms, key management, and the importance of securing private keys. 3. Forensics: This part of the report is dedicated to digital forensics, where the tasks involve analysing metadata, decoding hidden messages in files, and tracing the origins of certain actions. It highlights the importance of forensic techniques in cybersecurity for tracing malicious activities and recovering sensitive information. 4. Vulnerabilities: The vulnerabilities section explores common weaknesses in systems, such as improper handling of data, insecure file storage, and exploitable code practices. It emphasises the significance of securing applications against common attacks, such as code injection and file manipulation.

An In-Depth Analysis of Network Security, Cry

I independently conducted a comprehensive research project analysing various Online Teaching-Learning (OTL) platforms, with a focus on architecture, performance, scalability, security, and privacy. The study explored how these systems are reshaping traditional learning approaches, especially during the COVID-19 pandemic. Key platforms like Moodle, Google Classroom, and Zoom were compared based on their support for synchronous and asynchronous learning, multimedia capabilities, and user adaptability. Additionally, I assessed the feasibility of implementing OTL systems in Nepal, considering infrastructure, accessibility, and technical proficiency challenges. My research provided actionable recommendations for selecting the most appropriate platform for diverse educational environments, with a focus on scalability, user experience, and security features.

Comparative Analysis of Online Teaching-Learn

Our team conducted a comprehensive analysis of Walmart's weekly sales data from February 2010 to October 2012, focusing on economic conditions, regional factors, and holiday effects. We leveraged Power BI, R-driven methods, and Microsoft Excel to identify patterns and correlations. Key findings highlight the significant impact of holidays on sales, with recommendations for optimising inventory and promotional strategies during these periods. Additionally, store-specific metrics, such as sales performance, consistency, and elasticity, were analysed, providing insights for Walmart to enhance operational efficiency, align strategies with consumer behaviour, and maintain a competitive edge in the retail sector.

Walmart Sales Insights: Evaluating Trends and

This project involved the comprehensive setup, management, and automation of cloud resources on Amazon Web Services (AWS). The tasks began with the installation of UTM and Kali Linux on a macOS environment, facilitating the creation of a virtual machine for secure testing and deployment. Key activities included configuring the AWS Command Line Interface (CLI) with secure access credentials, creating and managing EC2 instances, and handling S3 bucket operations for cloud storage. The project also involved developing Python scripts using the Boto3 SDK to automate various AWS services, enhancing efficiency and reducing manual intervention. Data security was a major focus, with implementations of encryption techniques using AWS Key Management Service (KMS) and custom AES encryption. This ensured that sensitive data was securely stored and transferred within the AWS ecosystem. Additionally, the project covered the setup and management of DynamoDB, a NoSQL database service, to store and retrieve data efficiently.

AWS Cloud Implementation and Automation Proje

Project Overview: Collaborated on a data warehousing project focusing on designing and querying a graph database using Neo4j. The project revolved around analysing data from the 2014 FIFA World Cup, where we structured a property graph model to represent players, clubs, and countries as distinct nodes. We implemented an ETL process in Python to extract, transform, and load data into the graph database, ensuring data integrity and scalability. Key Contributions: • Designed a graph database schema that supports complex queries and efficient data retrieval. • Developed Cypher queries to answer specific questions related to player statistics, club affiliations, and national team performances. • Compared graph database capabilities with relational databases, highlighting the efficiency of graph structures for recursive and complex queries. • Explored practical applications of graph data science in finance, supply chain, and life sciences.