George Akai

Role: Team developer · Timeline: Summer (Cal Poly DXHub AI Summer Camp) · Context: Cal Poly DXHub AI Summer Camp; CSU-wide collaboration

During the Cal Poly DXHub AI Summer Camp, my team and I developed a solution to help Humboldt University’s research department navigate their website more efficiently, addressing issues of clickrage caused by multiple links on a single page. We designed a responsive interface using Streamlit and CSS, paired with a modern UI/UX that incorporated custom colors inspired by the Humboldt University website.

To enhance functionality, we implemented Retrieval Augmented Generation (RAG) for dynamic content generation and leveraged Amazon Bedrock for its RetrieveAndGenerate feature. Data was stored in Amazon S3 Buckets and indexed with an S3 vector database, while Amazon Knowledge Base features and a Semantic Chunking Strategy improved information retrieval and contextual understanding. For advanced Natural Language Processing (NLP), we used Titan V2 Embeddings and Claude Sonnet, supported by System Prompt Engineering to optimize AI model outputs. For security, we implemented Guardrails to monitor and control AI behavior, ensuring safe, accurate and reliable interactions.

Throughout the project, we documented our progress in GitHub, ensuring proper version control and adherence to good development practices. Greatest of all we experienced was the privilege of meeting and collaborating with excellent students from across the CSU system, each bringing unique professional backgrounds. Together, we built powerful AI solutions that addressed real-world problems. Huge thanks to Ryan Matteson, Nick Osterbur, Darren Kraker, and all the other admins for creating such a transformative summer camp that made a lasting impact on everyone who participated.

Key contributions

• Used Amazon S3 Buckets, S3 vector database, and Amazon Knowledge Base for retrieval quality and scale.
• Applied Semantic Chunking, Titan V2 Embeddings, and Claude Sonnet for stronger contextual responses.
• Implemented Guardrails and prompt controls for safe, accurate, reliable AI behavior.
• Documented progress in GitHub; collaborated with students across the CSU system.

Role: Lead developer · Timeline: Jan 2026 – Present · Context: CIS Capstone, California State University, Chico

For my Computer Information Systems Capstone project, I am developing a full-stack web application designed to streamline digital forensic investigations. The platform automates the extraction and parsing of artifacts from forensic acquisitions, enabling analysts to quickly surface relevant evidence.

Parsed data is securely stored in an AWS S3 bucket and leveraged to power an AI-driven analysis agent capable of generating detailed timelines and investigative reports. The application also includes an AI chatbot that uses Retrieval-Augmented Generation (RAG) to answer analyst queries and provide context-aware insights based on the collected artifacts.

This project integrates AWS cloud services, automation, and modern AI techniques to enhance the speed, accuracy, and accessibility of forensic analysis workflows.

Key contributions

• Automated artifact extraction and parsing from forensic acquisitions (PCAPs, registry hives).
• Architected secure storage pipeline using AWS S3 for parsed forensic data.
• Built AI agent with RAG on AWS Bedrock for timeline generation and investigative reporting.
• Implemented AI chatbot for context-aware analyst query answering from artifact corpus.

Role: Full-stack developer · Timeline: Ongoing · Context: Family cosmetics and beauty business · kenrishcollection.onrender.com

Kenrish Collection is a responsive, full-stack Inventory Management System (IMS) designed for boutique retail businesses specializing in cosmetics, apparel, and fashion accessories. The platform streamlines back-office workflows by giving store administrators real-time visibility into stock levels, operational costs, product pricing, and financial health.

Core admin features include real-time inventory tracking across diverse product categories with automated low-stock warnings, a Dynamic Valuation Engine that calculates total inventory value per item based on real-time cost vs. retail pricing, and integrated stock and sales management modules for updating wholesale costs, adjusting consumer pricing, and recording transactions. A historical Ledger and Revenue Logging database is linked to point-of-sale workflows for accurate profit margin calculation and financial reporting.

The platform includes a Retrieval-Augmented Generation (RAG) chatbot controlled by a model harness using OpenRouter models. It uses NLP to answer customer questions about services and products offered by the business. Authentication is secured via JWT tokens, and the UI supports dark and light mode with a clean, mobile-friendly tabbed dashboard built for quick navigation and rapid data entry on the retail floor.

Key contributions

• Full-stack IMS: React frontend, Django REST backend, deployed on Render.
• Dynamic Valuation Engine with real-time cost vs. retail pricing and automated low-stock alerts.
• RAG chatbot using OpenRouter models and a model harness for controlled NLP-driven customer Q&A.
• JWT-secured auth; dark/light mode; mobile-friendly tabbed admin dashboard.

Screenshots

Role: Competitor · Timeline: Fall 2024 · Context: National Cyber League

During my participation in the National Cyber League, I tackled a wide range of cybersecurity challenges that tested both technical skill and problem-solving ability. I began with OSINT investigations uncovering publicly available information. From there, I moved into cryptography, applying encryption and decryption techniques to solve complex puzzles or crack encrypted passwords. I also analyzed network traffic to identify anomalies and potential vulnerabilities, while system enumeration exercises taught me how to detect open ports, services, and exploitable configurations. Finally, I engaged in password cracking, working with hashed data to retrieve flags and gain deeper insight into real-world attack scenarios. Together, these experiences strengthened my ability to think critically, adapt quickly, and apply red-team cybersecurity principles in practical contexts.

Role: Learner · Timeline: 2023-2024 · Context: Google / Coursera

Through the Google Cybersecurity Certificate program, I built a strong foundation in security principles, including the CIA triad and the NIST Cybersecurity Framework. I gained hands-on experience working with Linux and SQL, while also learning SOC workflows to understand how security operations are managed in real-world environments. Using Wireshark, I learned to analyze network traffic and identify potential threats, and with Suricata I conducted vulnerability assessments while applying risk and asset management strategies. I also explored security automation by writing Python scripts to streamline repetitive tasks, and studied the importance of playbooks in guiding effective incident response. Beyond the technical skills, the program emphasized how to make an immediate impact as a cybersecurity analyst, integrating seamlessly into a team and contributing from the very first day on the job.

Role: Information Security Analyst Assistant · Timeline: Feb–Jun 2025 · Context: Information Security Office, Chico State

Working in the Chico State Information Security Office, I supported a broad range of security operations under the direction of the CISSP-qualified Information Security Officer and in close collaboration with the Risk Management Officer.

For vulnerability management, I generated and analyzed Qualys VMDR vulnerability reports, tracked remediation progress, and coordinated follow-up with campus departments on patching compliance and end-of-life servers — improving team efficiency by 15% and reducing outstanding high-risk vulnerabilities. I also supported incident-remediation discussions and documented next steps in the CMDB, enabling faster resolution of campus security concerns by 15%.

For identity and access management, I led annual user access reviews to validate RBAC permissions across campus applications and data systems in all departments, ensuring proper access control and reducing unauthorized access risk. I managed ticket workflows in collaboration with campus MPPs and distributed certification requests via Adobe Sign.

For compliance and awareness, I authored knowledge-base and procedural documentation for the annual sensitive data inventory survey. I prepared DMARC email-conformance communications for technical, managerial, and non-technical stakeholders, and tracked security-awareness training completion via CSU Learn, coordinating follow-up with users and improving campus-wide participation.

For threat intelligence, I monitored CISA advisories, CVEs, and indicators of compromise to support proactive threat detection, enabling early identification of potential threats and faster response.

Key contributions

• Qualys VMDR vulnerability reporting and remediation tracking — 15% efficiency improvement.
• Led RBAC user access reviews across all campus departments; documented findings in CMDB.
• Authored DMARC compliance communications for technical and non-technical stakeholders.
• Monitored CISA advisories, CVEs, and IOCs for proactive threat detection.
• Tracked security-awareness training (CSU Learn) and coordinated follow-up, increasing participation.
• Collaborated with CISSP ISO and Risk Management Officer on vulnerability exceptions and risk assessment.

Role: Author / technical writer · Timeline: 2025 (annual survey cycle) · Context: Chico State ISEC

While working in the Information Security Department, I was assigned to create a KnowledgeBase article that would guide Chico State departments in preparing for the annual Sensitive Data Inventory Survey. I authored the article, which was published in the ISEC documentation on the school’s support page, ensuring that departments had clear instructions and resources to complete the survey effectively. The article emphasized the importance of protecting Level 1 data (CSU’s classification for highly sensitive information) and promoted compliance awareness by explaining how the survey supported adherence to CSU policy. In doing so, I contributed to broader security initiatives by providing clear documentation and communication that reinforced the university’s commitment to data protection and compliance.

Role: Computer Deployment Technician · Timeline: Jun–Aug 2025 · Context: IT Support Services, Chico State

As a Computer Deployment Technician with IT Support Services at Chico State, I was responsible for the campus PC-as-a-Service rollout — deploying, imaging, and configuring over 300 Windows 11 PCs across campus labs and offices using Microsoft Deployment Toolkit (MDT).

I managed deployment tickets and workflow documentation in Team Dynamix, ensuring all tickets were resolved within service-level agreements. After machines were deployed and users connected to their accounts, I supported end-user needs and service delivery — improving service delivery by 20%. I also followed Chico State e-waste procedures to retire legacy equipment safely, achieving full compliance with university environmental policies.

Key contributions

• Deployed, imaged, and configured 300+ Windows 11 PCs using Microsoft Deployment Toolkit — two weeks ahead of schedule.
• Managed deployment tickets and SLA-compliant workflow documentation in Team Dynamix.
• Supported end-user onboarding and account setup, improving service delivery by 20%.
• Retired legacy hardware per Chico State e-waste procedures, achieving full environmental compliance.

Role: Team developer · Timeline: Course midterm · Context: Chico State software development course

As part of a software development class midterm project, my team and I set out to build a fully functional chess game board using Django. The goal was to explore how Django handles backend requests compared to other stacks that rely more heavily on client-side rendering. On the frontend, we designed templates with Bootstrap, CSS, HTML, JavaScript, and JQuery, ensuring a responsive and visually appealing interface. The backend was powered by Django and a SQLite3 database, where we implemented models, views, and URLs to render the board accurately. These models provided the foundation of the project, while the views acted as handlers that connected templates and URLs seamlessly. Movement of chess pieces was achieved through a combination of JavaScript and Django backend requests, allowing for dynamic interaction across the board. To safeguard user accounts, we applied security best practices for form validation, signup, login, and logout functionality. Finally, we documented our progress and maintained version control using GitHub, reinforcing good development practices. This project was both challenging and rewarding, offering hands-on experience in blending frontend design with backend logic while building something fun and interactive.

Key contributions (summary)

• Implemented Django models, views, and URLs with SQLite3 to render game behavior.
• Connected JavaScript-driven piece movement with backend handling for dynamic interaction.
• Applied security best practices for form validation, signup, login, and logout.
• Maintained version control in GitHub.

Role: Developer · Timeline: Early version (prior iteration) · Context: Family business storefront

For this project, I set out to build a web application for my sister’s cosmetics and beauty shop, aiming to solve the challenge of visibility and give her business a stronger online presence. The goal was to make her products and services easily accessible to customers through both mobile devices and computers. I designed the site with a responsive, mobile-first approach using HTML, CSS, and Bootstrap, with a Node.js and Express.js backend, ensuring that the layout adapted seamlessly across different screen sizes. To enhance the user experience, I implemented a modern UI/UX with a custom color palette and smooth navigation, creating an inviting and intuitive interface. The platform showcased her offerings clearly, while also incorporating security measures such as password encryption to protect user accounts. This early Node.js version of the Kenrish Collection website laid the foundation for a secure, user-friendly digital storefront that brought her business into the spotlight.