résumé

Nathan Romeo

Artificial intelligence & machine learning student

Email
nromeo8@gmail.com
Phone
0435 269 704
LOCATION
Queensland, Australia
Study
University of Southern Queensland
Degree
information technology
majors
· AI & Data Analytics
· Software Development

summary

Backend-focused IT graduate specialising in AI-assisted systems and structured software design. Experienced in building modular full-stack applications using FastAPI, PostgreSQL, React, and containerised deployment workflows. Strong interest in intelligent system architecture, prompt engineering and production-ready AI integration.

Bringing over ten years of leadership experience running a production company, managing compliance, clients, and full project lifecycles. Known for translating complex requirements into practical, scalable systems grounded in real operational needs.

Technical skills

Languages
Python, JavaScript, R, C++, HTML/CSS
Backend Development:
FastAPI, Pydantic, SQLAlchemy
Frontend Development:
React, TypeScript, React Router
Database:
PostgreSQL, MongoDB, MySQL, relational schema design, JSON
DevOps & Deployment:
Docker, container networking, environment configuration
Version Control:
Git, feature branching, merge conflict resolution

Professional Skills

  • Systems thinking & modular architecture design

  • Process automation & workflow optimisation
  • Stakeholder communication
  • Requirements analysis
  • Multicultural collaboration
  • Project lifecycle management
  • Regulatory-aware system design
  • Analytical problem solving
  • Independent learning
  • Ethical AI awareness
  • Deployment & environment configuration
  • Performance optimisation
  • Functional collaboration

Technical COMPETENCIES

  • Machine Learning

  • LLM Integration

  • Backend API Development

  • Model Evaluation

  • Data Engineering

  • Containerised Deployment

  • Prompt Engineering

  • Structured Data Pipelines

  • Systems testing

Experience & Education

Experience

2016-PRESENT

Managing Director, Video Producer & Technical Lead

Foxtail Productions · 2016 - Present
I manage the operations of Foxtail Productions, working across both creative and business functions. My role covers the full production lifecycle, from concept and scripting through to directing, filming, and post-production, delivering content for television, digital, and social platforms.
Alongside production, I work directly with clients on strategy, copywriting, and campaign direction, while also managing quoting, invoicing, and project coordination. I have extensive experience delivering broadcast-compliant media and over 6 years of drone operation under a RePL licence.

2015-PRESENT

Camera operator

Toowoomba Speedway · 2015 - Present
I film live racing events in a fast-paced, high-pressure environment, capturing footage for live-streamed broadcasts. This role requires quick decision-making, strong situational awareness, and the ability to consistently deliver usable footage under dynamic conditions. I am also responsible for aspects of AV setup and configuration, including professional cable management and ensuring reliable operation during live events.

2016-2017

Video Editor & Camera Operator

Channel 9 Toowoomba · 2016 - 2017
At Channel 9 Toowoomba, I worked closely with journalists to film and edit news stories for broadcast. This role provided hands-on experience in the fast turnaround of news production, as well as a strong understanding of broadcast standards, storytelling and working within tight deadlines.

EDUCATION

2023-present

University of Southern Queensland - TOOWOOMBA

Bachelor of Information Technology
AI & Data Analytics • Software Development
University of Southern Queensland • 2023 – Present

• Focused on backend system architecture, AI integration, prompt engineering, database design and full-stack application development.
• Designed and implemented REST APIs using FastAPI with structured request/response validation via Pydantic.
• Developed relational database schemas using PostgreSQL and SQLAlchemy, including foreign key relationships and JSON data storage.
• Applied machine learning and data analysis techniques in Python and R for structured data processing.
• Built and deployed containerised applications using Docker, including environment configuration and container networking.
• Completed an industry-facing capstone project (CareShield AI), developing an AI-assisted documentation system integrating speech-to-text, multilingual handling and persistent relational storage.
• Applied modular software design principles to ensure scalability and maintainability.

2025-2026

Darmstadt University of Applied Sciences - GERMANY

Digital Transformation & Intercultural Communication
Darmstadt University of Applied Sciences • 2025 • Semester exchange

• Completed an intensive international program examining digital transformation strategies in modern organisations.
• Analysed how emerging technologies (including AI) impact governance and regulatory compliance.
• Collaborated within multicultural teams to evaluate global technology adoption challenges.
• Produced a structured reflective portfolio analysing intercultural communication and digital implementation.

Projects

CareShield AI : Full Stack AI Documentation System

Designed and developed an AI-assisted conversational case note system for an aged care provider.

  • Built modular FastAPI backend with structured AI pipeline (translation, follow-up logic, structured extraction)
  • Integrated Whisper for speech-to-text processing
  • Implemented multilingual detection and contextual translation
  • Enforced structured JSON extraction and validation for deterministic database storage
  • Designed PostgreSQL schema for session-based conversational persistence
  • Containerised backend using Docker for deployment portability
  • Built React-based conversational UI

Technologies: Python, FastAPI, PostgreSQL, React, Docker, Whisper, Ollama (Qwen 2.5)

Nathan AI: Full Stack Portfolio Chat Assistant

Developed a personalised AI chat bot assistant tailored to my personality for deployment on a live website. 

  • Designed a Python FastAPI service with integrated OpenAI LLM gateway via Groq using httpx
  • Grounded replies with version-controlled persona/knowledge Markdown bundled into the backend image
  • Configured CORS for multiple production origins (Vercel, custom domain WordPress embed)
  • Containerised the API with Docker and deployed to Render
  • Responsive React + Vite UI with streaming-style loading states, fetch-based integration and Git-backed continuous deployment on Vercel
  • Embedded the assistant on a production WordPress site

Technologies: Python, FastAPI, React, Vite, Docker, Groq API, Render, Vercel, httpx

Rocky Creek Homestead: Booking Website

Designed and developed a full-service promotional and booking website for a real-world accommodation business.

  • Led complete end-to-end system design, including UX flow, booking integration, hosting configuration, and deployment
  • Implemented and configured booking workflow to support real-world availability management and enquiry handling
  • Managed domain migration, DNS updates, and hosting environment configuration
  • Structured user journey to reduce friction from landing page through to booking submission
  • Optimised site performance using LiteSpeed Cache and QUIC.cloud CDN
  • Configured caching layers, image optimisation, and server-side performance tuning
  • Diagnosed and resolved load-time bottlenecks caused by high media usage
  • Balanced SEO metrics with practical real-world responsiveness

Technologies: WordPress, Elementor, LiteSpeed Cache, QUIC.cloud CDN, CSS, DNS & Hosting Configuration

Video Export Workflow Automation: ExtendScript System

Designed and implemented an automation system to streamline the delivery of broadcast-compliant television commercials.

The production workflow required multiple export formats per commercial, each with strict technical specifications and naming conventions. Manual processing introduced inefficiencies and risk of human error across deliverables.

  • Developed a custom automation solution using Adobe ExtendScript to batch-process multi-format exports
  • Encoded predefined broadcast specifications directly into the script to enforce compliance standards
  • Standardised file naming conventions and output configurations across all deliverables
  • Reduced repetitive manual export steps to a single automated operation
  • Minimised risk of human error in high-pressure, deadline-driven production environments
  • Improved turnaround time for broadcast submissions

Technologies: Adobe ExtendScript (JavaScript-based scripting), Premiere Pro automation, broadcast compliance standards

Depth of Anaesthesia Index: Hybrid Machine Learning Model

Developed a hybrid machine learning model to estimate depth of anaesthesia (DoA), aiming to improve reliability compared to the traditional Bispectral Index (BIS).

  • Designed and implemented a supervised learning model using a linear Support Vector Machine (SVM) to generate continuous DoA predictions aligned with BIS reference values
  • Applied K-means clustering to identify underlying anaesthetic state patterns without labelled supervision
  • Engineered a hybrid modelling approach combining supervised and unsupervised outputs using a weighted integration parameter (α)
  • Performed feature selection based on correlation analysis, prioritising high-signal predictors and eliminating redundant inputs
  • Tuned weighting parameter using validation data to minimise prediction error
  • Evaluated model performance using statistical metrics including correlation (r), RMSE, and R²

Performance Results:

  • Correlation (r): 0.905
  • RMSE: 8.06
  • R²: 0.792

Technologies: Python, scikit-learn, SVM, K-means clustering, statistical validation techniques

ABS Census Form: Front-End Form Design (Mock Implementation)

Developed a census-style data collection form using HTML and CSS, focusing on usability, accessibility and realistic input behaviour.

  • Designed clean, structured layout with logically grouped inputs and consistent visual hierarchy
  • Implemented semantic HTML structure with properly associated labels for accessibility compliance
  • Applied client-side validation rules to balance data integrity with realistic user input formats
  • Structured form flow to minimise friction and guide users through multi-section data entry
  • Replicated real-world government-style data collection principles in a simplified interface

Technologies: HTML5, CSS3, form validation logic

Download Resume
Download Transcript