Nicholas Aplin Email & Phone Number

Adelaide, South Australia, Australia

Key Achievements:- Clustering Algorithm Development: Developed and implemented a clustering algorithm to group assets for predictive analytics, leading to AI model development for failure prediction. Conducted extensive corrosion data analysis, optimising inspection schedules and reducing work by 15% without increasing failure risk.- Corrosion Data Integration: Integrated internal inspection data with external geographic, demographic, and meteorological datasets, applying AI algorithms in Python (Random Forests, KNN, K-Means) to analyse corrosion patterns and predict asset failures.- Computer Vision and OCR: Applied computer vision detection and optical character recognition techniques to enhance PDF document extraction, significantly improving table extraction. Enabled the searchability of unstructured data using Large Language Models.- Chat Interface Enhancement: Enhanced an open-source chat interface built with Svelte and Python by enabling non-standard LLM APIs, integrating improved document extraction, and improving document retrieval using RAG techniques including HyDE and LLM reranking. Utilised Azure DevOps for version control and software development.Additional Responsibilities:Ad hoc data analysis including geospatial analysis and data modelling.Keeping up with and reporting on the latest advancements in AI and data science.

Adelaide, South Australia, Australia

In this role, I have leveraged computer vision, deep learning, and AI technologies to enhance operational efficiencies and predictive capabilities within the organisation.Key Projects- Document Analysis with Computer Vision: Implemented computer vision to analyse documents, integrating with a vector database to improve searchability and provide contextual insights for an internal procedures chatbot.- Asset Type Identification: Applied Azure Custom Vision and Python to identify asset types from inspection photos, recording tags and probabilities for replacement assessments, reducing costs.- LLM-powered Chatbot Development: Developed an LLM-powered chatbot POC, integrating Azure OpenAI API with added professionalism guardrails.- Predictive Analytics and AI Model Development: Engineered an asset clustering algorithm for predictive analytics and AI model development.- Image Classification Model Training: Deployed advanced vision labelling software for image classification model training using segmentation techniques.- Automated Asset Inspection: Utilised deep learning to automate asset inspection, reallocating resources to higher-value projects.Additional Responsibilities:- Produced general analytics reports and explored innovative solutions for legacy business processes.- Trained computer vision models using Azure Custom Vision and Python, identifying specific asset types and recording probability scores for replacement.- Integrated internal inspection data with external datasets for extensive corrosion analysis using AI algorithms like Random Forests, KNN, and K-Means clustering to predict future failures.- Engaged with stakeholders to assess and quantify the value of collected data points during asset inspections.- Developed a computer vision model using PyTorch, enhanced with segmentation and transfer learning to increase accuracy and reduce training time.

Adelaide, South Australia, Australia

During this internship I undertook 2 projects:- analysing data a dataset and creating a presentation based around the feasibility of a new software- Use python and NLP packages such as NLTK, Gensim and Spacey to analyse long text safety reports to find new insights into what causes safety incidents using a combination of topic analysis and word frequency analysis

Adelaide, South Australia, Australia

Bachelor Of Mathematical Sciences (Advanced), 5.5 Gpa

Activities and Societies: Adelaide university maths society

High School Diploma, 98.15