Skip to content
View Neill-Erasmus's full-sized avatar
😎
😎
14 followers · 26 following

Achievements

Achievement: YOLOAchievement: QuickdrawAchievement: Pull Sharkx2

Achievements

Achievement: YOLOAchievement: QuickdrawAchievement: Pull Sharkx2

Highlights

  • Pro

Block or report Neill-Erasmus

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Maximum 250 characters. Please don’t include any personal information such as legal names or email addresses. Markdown is supported. This note will only be visible to you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
Neill-Erasmus/README.md

👋 Neill Erasmus

Python PyTorch TensorFlow Scikit-Learn Keras

🎯 AI/ML Engineer | Honours Computer Science Student

I'm a Computer Science Honours student with a passion for Artificial Intelligence and Machine Learning. I love developing AI solutions that solve real-world problems. My work bridges the gap between research and practical applications.

🔹 Current Focus

  • Deep learning and neural networks (PyTorch / TensorFlow)
  • Machine learning pipelines & data preprocessing
  • Expanding my portfolio with practical, real-world projects

🔹 Featured Projects

🌾 Sugarcane Leaf Disease Detection

AI-powered agricultural solution for disease diagnosis

  • Tech: Python, PyTorch, Convolutional Neural Networks (CNNs), Docker
  • Impact: High-accuracy multi-class image classification for agricultural applications
  • Highlights:
    • Production-ready API with containerized deployment
    • Real-time inference optimization
    • Comprehensive documentation and usage examples

🔍 Agentic Research Assistant

Planner-driven multi-agent system for automated research and report generation

  • Tech: Python, Ollama, Multi-Agent Orchestration, Docker
  • Impact: Transforms raw research queries into structured, cited reports with LLM-backed summarisation and fact-checking
  • Highlights:
    • Four specialised agents (Search, Summariser, Fact-Checker, Citation) with dependency-aware pipeline execution
    • Multi-source web retrieval with deduplication, relevance ranking, and source-diversity balancing
    • Deterministic fallback behaviour ensures reliability when the LLM is unavailable

🧠 The Networked Agent

Local AI reasoning framework combining Graph of Thought and GraphRAG

  • Tech: Python, Ollama, Graph of Thought (GoT), GraphRAG, Docker
  • Impact: Enables complex multi-step reasoning over domain knowledge entirely on-device, with no hosted API dependency
  • Highlights:
    • Beam search over candidate reasoning paths with hybrid heuristic and LLM-based scoring
    • Multi-hop knowledge graph retrieval with vector similarity search and evidence tracking
    • Per-query visualisation artifacts (HTML, JSON, DOT) for full reasoning transparency

️ Tech Stack

Languages:
Python C# Java SQL

AI/ML & Deep Learning:
PyTorch TensorFlow Scikit-Learn

Data & Visualization:
Pandas NumPy SciPy Matplotlib

DevOps & Tools:
Docker Git GitHub

Let's Connect

I'm open to collaborations, discussions about AI/ML, and exciting opportunities!

LinkedIn Email GitHub

Pinned Loading

  1. the-networked-agent the-networked-agent Public

    A networked AI agent implementing Graph of Thought (GoT) and GraphRAG for non-linear, human-like reasoning. Built for privacy and performance using local models via Ollama.

    Python

  2. agentic-research-assistant agentic-research-assistant Public

    Planner-driven multi-agent research assistant for local Ollama models with web retrieval, summarization, fact-check cues, and APA citations.

    Python 1

  3. sugarcane-leaf-disease-detection sugarcane-leaf-disease-detection Public

    End-to-end deep learning system for sugarcane leaf disease detection using PyTorch, ResNet-50, FastAPI, and Docker.

    Python 2