AWS, GCP, Java, JavaScript, TypeScript, C++, Rust, Solidity, Jenkins, Ansible, Grafana, OWASP, Trivy, Sonarqube, Terraform, Python, Git, AWS VPC, EC2, S3, EKS, ECS, ECR, Lambda, AWS IAM, Azure, Cloud Computing, CI/CD, Prometheus, Shell, Kubernetes, OOP, Compute Engine, Google Cloud VPC, GCP IAM, DevOps, DevSecOps, Blockchain

Software Engineer @ Web3 Builders Alliance (February 2023 - Present)

• Building BlockScraper, a blockchain product and marketing analytics project with NextJS, Vercel and machine learning.
• Architecting and building Solana Arduino SDK with C++ and DevSecOps best practices.
• Engineered data in Grizzlython for the SolSync project with TypeScript and researched public cloud use on AWS and GCP and provided a proof of concept (PoC) on machine learning.
• Working with DevOps pipelines with emphasis on security with Sonarqube, Trivy, OWASP, Jenkins, Grafana, Kubernetes, Docker and Prometheus, reducing product delivery by 30% and increased security by 25% with deployments.

Software Engineer @ Google App and Game Academy (December 2022 - Present)

• Built and launched a mobile NFT marketplace app on the Ethereum network using Flutter and Hardhat.
• Worked on ZKML in the Ethereum blockchain.
• Developed a food app, a TikTok clone, and a Twitter clone with Docker and Appwrite.
• Led a team of 3 software engineers in a multidisciplinary setting.

Software Engineering Intern @ Patika.dev (May 2021 - January 2023)

• Engineered software on EVM blockchains and Solana with Rust, Solidity and TypeScript.
• Developed and deployed real-life blockchain projects in Rust and Solidity.

This project demonstrates my expertise in deploying and securing cloud-based applications using a DevSecOps approach. I built a Netflix clone application and implemented a CI/CD pipeline with Jenkins for automated deployment and scaling. The pipeline integrates security best practices throughout the development lifecycle, including vulnerability scanning with Trivy and SonarQube for code and containerized images. Additionally, Prometheus and Grafana provide real-time monitoring and insights into the application's health and performance. This project showcases my ability to leverage DevOps tools and practices to build, secure, and deliver reliable cloud-based applications.

This project streamlines Zomato clone deployment with a secure DevSecOps CI/CD pipeline. Utilizing tools like Jenkins, Docker, SonarQube, OWASP, and Trivy, the pipeline automates code checks, vulnerability scans, and deployments. From cleaning the workspace to containerization and final image scans, every stage ensures security and reliability. This automation led to a 60% faster deployment time, improved application stability through continuous integration, and enhanced security through proactive vulnerability detection. This project showcases my expertise in building and deploying secure, cloud-based applications.

Project Link

Spearheaded the transformation of a monolithic application into a robust and scalable microservice architecture on AWS EKS. This ambitious project involved orchestrating the decomposition of a monolithic codebase into 12 microservices. Leveraging cloud-native technologies, I implemented robust automation and observability measures to ensure seamless operation and performance insights. This strategic move empowers the application with enhanced scalability, high availability, and agile development capabilities, paving the way for future growth and innovation.

Successfully spearheaded the migration of a monolithic application to a scalable and resilient microservices architecture on AWS EKS. This strategic initiative involved decomposing the application into 6 independent services, each owned by dedicated development teams. Leveraging cloud-native technologies like Kubernetes and containerization, I implemented automated deployments, continuous integration/continuous delivery (CI/CD) pipelines, and comprehensive monitoring and observability tools. This transformation resulted in significant improvements in scalability, agility, and fault tolerance, enabling the application to handle increased traffic and adapt to evolving business needs.

This project showcases a secure, automated StreamSafe Kids deployed on AWS using DevSecOps principles. Tools like Docker, Jenkins, SonarQube, and Docker Scout seamlessly weave security into every stage, from static code analysis to container scans. Enjoy efficient deployments with Kubernetes orchestration, all on a robust security foundation.

Project Link

Project Link

Addressing the critical issue of coccidiosis detection in poultry, I spearheaded the development of a deep learning model for accurate classification. This project entailed:

Data Acquisition: A comprehensive dataset of 350 fecal images was meticulously curated from diverse sources, ensuring balanced representation of both healthy and diseased chickens.

Preprocessing and Augmentation: Rigorous image preprocessing with resizing, grayscale conversion, and augmentation techniques (flipping, rotation) enhanced data robustness.

Model Design and Training: Utilizing the Keras library, a convolutional neural network architecture was implemented and trained on the preprocessed data, achieving an impressive 83% accuracy on the held-out test set.

Cloud Deployment: For seamless accessibility and scalability, the model was deployed on both AWS and Azure platforms. A dedicated Jupyter notebook facilitates further experimentation and prediction capabilities.

Dataset Balancing: Addressing the inherent imbalance in chicken health representation within the data through oversampling techniques.

Computational Efficiency: Leveraging cloud GPUs to expedite model training and optimize resource utilization.

Production-Ready Deployment: Navigating the complexities of deploying a deep learning model to production environments, opting for cloud platforms for robust accessibility and scalability.

Impact and Significance: This project successfully delivers a valuable tool for poultry farmers and veterinarians, empowering them with accurate and efficient coccidiosis detection. The cloud-based deployment ensures widespread accessibility and paves the way for further integration within poultry management systems.

Project Link

Leveraging the state-of-the-art YOLOv8 deep learning model, I built an application for accurate cell segmentation in medical and genetic images. This project involved:

Model Selection: Recognizing YOLOv8's suitability for both object detection and segmentation tasks, I opted for it as the core of the application. Dataset Preparation: A crucial step involved curating a labeled dataset of medical and genetic images depicting various cell types. To ensure compatibility with YOLOv8, the labels were converted to the appropriate format.

Model Training: Utilizing a cloud platform to harness sufficient computational resources, I trained the YOLOv8 model on the prepared dataset. This rigorous training process optimized the model's cell segmentation accuracy.

Application Development: Building upon the trained model, I developed a user-friendly application in Python, employing the Flask framework. This application allows users to seamlessly upload new images and receive precise cell segmentation results.

Deployment and Accessibility: For broader impact, the application was deployed to Azure, enabling easy access for researchers and clinicians to leverage its cell segmentation capabilities for their own studies and analyses.

Data Acquisition and Labeling: Gathering and accurately labeling a diverse dataset of medical and genetic images posed a significant challenge. Collaborating with domain experts and employing efficient labeling techniques ensured high-quality data for model training.

Computational Demands: Training a deep learning model like YOLOv8 requires substantial computational resources. Utilizing cloud platforms provided the necessary scalability and efficiency to complete the training process successfully.

Deployment and User Interface: Ensuring seamless cloud deployment and designing a user-friendly interface were crucial for maximizing the application's accessibility and usability for researchers and clinicians.

Project Link

Successfully led the development and deployment of a cutting-edge waste detection and classification system utilizing YOLOv5 deep learning technology. This project addressed the critical challenge of waste management by:

Model Selection and Training: Recognizing YOLOv5's proficiency in object detection, I implemented and trained a custom model on a curated dataset of labeled waste images, achieving high accuracy in identifying various waste categories.

Streamlined Workflow: The system leverages Docker containers for efficient deployment and scalability, enabling seamless integration with existing waste management infrastructure.

Real-time Processing: Utilizing optimized algorithms, the system facilitates real-time waste detection and classification, allowing for immediate sorting and processing of waste streams.

Data Visualization and Analysis: A dedicated dashboard provides comprehensive insights into waste composition and collection patterns, empowering informed decision-making for optimized waste management strategies.

Data Acquisition and Labeling: Assembling a diverse and accurately labeled dataset of waste images posed a significant challenge. Collaborating with waste management facilities and employing efficient labeling techniques ensured high-quality data for model training.

Model Optimization: Balancing accuracy with computational efficiency was crucial for real-time deployment. Through careful hyperparameter tuning and architecture optimization, the model achieved optimal performance on edge devices.

Integration and Scalability: Integrating the system with existing waste management infrastructure required careful consideration of compatibility and data security. The Docker-based containerization approach facilitated seamless deployment and scalability across diverse environments.

Demo Video

SolSync represents a pioneering solution to the challenge of obtaining off-chain data from blockchains, addressing the limitations of relying solely on on-chain data. Leveraging the Twitter API to aggregate diverse tweets, our meticulous research and development process unfolded during the Solana Grizzlython, resulting in a three-part project encompassing a front-end application, the SolSync Program, and a black box program. Tailored for Solana dApps, SolSync facilitates secure and efficient access to off-chain data, employing a Solana blockchain-deployed black box to store and manage data securely. My role encompassed research, development, and deployment, extending to a comprehensive proof of concept for cloud infrastructure, cross-chain integration, and data infrastructure, where AWS emerged as the most efficient choice. SolSync's operational intricacies are visually depicted in the accompanying diagram, showcasing a robust system that seamlessly fetches, stores, and protects off-chain data. This project underscores my expertise in blockchain development, cloud infrastructure, and cross-chain solutions, positioning me as a valuable contributor to pioneering projects in the blockchain space.

Google Project Management