Previous Events
Explore the milestones and achievements of EpochDev through our past hackathons and workshops.
ποΈ The Eyes of AI β Object Detection
- π Date: December 2025
- π Location: National College of Ireland β Spencer Dock - Room 3.05
- π₯ Participants: 10
This workshop introduced attendees to object detection, a core computer vision task focused on locating and classifying objects within images. Participants explored how modern systems have evolved from traditional feature engineering to deep learningβbased approaches, with a strong emphasis on YOLO (You Only Look Once) for real-time detection. The session concluded with a guided, hands-on walkthrough of fine-tuning YOLO on a custom football dataset.
π Session Outline
- ποΈWhat is object detection & real-world use cases
- π§ From edge detection to neural networks (AlexNet β YOLO)
- π¦YOLO architecture, grids & bounding boxes
- βοΈFine-tuning YOLO on a football dataset
π Project Spotlight β The Eyes of AI
Attendees worked with a real object detection pipeline using YOLO, fine-tuned on a football dataset sourced via Roboflow. π GitHub: Object Detection
- βοΈStack: Python Β· YOLO Β· Roboflow
- π―Goal: understand & build a real-time object detector
π Resources
- ποΈ Yolo explanation
- π Slides
- π§ͺ Repository
-
πΎ Code Examples (quick start)
git clone https://github.com/Epochdev0/Object-Detection.gitcd Object-Detection- Create env:
python -m venv .venv && . .venv/bin/activate(macOS/Linux)
python -m venv .venv && .\\.venv\\Scripts\\activate(Windows) pip install -r requirements.txt- Run training or inference scripts as shown in the repo
π§ Computer Vision 101 β Part 2 (Extended Session)
- π Date: November 4th 2025
- π Location: National College of Ireland - Mayor Square Building Room 4.18
- π₯ Participants: 15
This extended session built on the foundations from Workshop #1 and transitioned into the world of neural networks and model optimisation. Participants explored hands-on examples of how modern computer vision systems are trained and stabilised using techniques like batch normalisation and dropout. The session also introduced integrating trained models into real-world applications, showing how deep learning pipelines can move from notebooks into production-ready solutions.
π Workshop Outline
- π§©Quick recap of convolutional layers & feature maps
- βοΈBatch Normalisation & Dropout β why they matter
- π€Building and testing small neural networks
- πModel integration and real-world deployment insights
π Project Spotlight β Computer Vision for AI 101 (Part 2)
The extended workshop used the official EpochDev Computer Vision repoβs Part 2 materials. π GitHub: Computer-Vision-for-AI-101 (Part 2)
- βοΈStack: Python + TensorFlow/Keras + OpenCV
- π―Goal: connect classical CV with neural network implementation
π Resources
- π§ͺ Part-2 Repository
- π LinkedIn Post
-
πΎ Code Examples (quick start)
git clone https://github.com/Epochdev0/Computer-vision-for-AI-101.gitcd "Computer-vision-for-AI-101/Part - 2"- Create env (pick one):
python -m venv .venv && . .venv/bin/activate(macOS/Linux)
python -m venv .venv && .\\.venv\\Scripts\\activate(Windows) pip install -r requirements.txt- Run demo notebooks or scripts to explore the models
π§ Computer Vision 101 Workshop
- π Date: October 16th 2025
- π Location: National College of Ireland - Spencer Dock Building Room 3.05
- π₯ Participants: 15
Our first workshop introduced the fundamentals of Computer Visionβpixels, features, and simple recognition pipelinesβwith a short live demo and hands-on practice.
π Workshop Outline
- π·Image basics: pixels, channels, resolution
- πEdges & features at a glance
- π§ From features to simple classifiers
- π»Mini live demo
π Project Spotlight β MNIST Classification
A clean baseline digit classifier on the MNIST dataset with training + evaluation scripts.
π GitHub: MNIST-Classification
- βοΈStack: Python + DL framework (see repo)
- π―Goal: quick, reproducible baseline for CV101
π Resources
- π Slides
- π§ͺ MNIST-Classification Repo
-
πΎ Code Examples (generic)
git clone https://github.com/Epochdev0/MNIST-Classification.gitcd MNIST-Classification- Create env (pick one):
python -m venv .venv && . .venv/bin/activate(macOS/Linux)
python -m venv .venv && .\\.venv\\Scripts\\activate(Windows) pip install -r requirements.txt- Run training (example):
python train.py - Evaluate/visualise (example):
python evaluate.pyor open the provided notebook
π NASA Space Apps Challenge 2025
- π Date: October 4th - 5th, 2025
- π Location: National College of Ireland, Dublin
- π₯ Teams: 2 (6-7 participants each)
The NASA Space Apps Challenge united two innovative EpochDev teams tackling real-world space data challenges. Participants built advanced AI tools for interpreting astronomical and biological datasets, combining creativity and technical excellence.
π Team 1: EpochDevNLP
Members: Amir Freer, Aditya Patil, Edward Chvainickas, Enis ΕimΕΔ°R, Harsha Varthan
- π Visual knowledge graphs connecting experiments, genes, and findings
- π¬ Natural-language chatbot with referenced answers
- π Interactive charts and filtering tools
π°οΈ Team 2: EpochDev ComV
Members: Anujin Ariunbold, Dhruv Waghela, Pawan Badsewal, Manisha Mani, Deepak Chandra Nallamothu, Temuulen Munkhtaivan, Johnpaul Ajuzie
- β‘ExoVision β classifies light-curve signals in seconds using Kepler-inspired features
- π Real-time single-signal and scalable batch mode (CSV upload)
- π§ Stack: XGBoost, FastAPI, React
- π Sustainability: optimized compute and cloud-friendly deployment
- π Roadmap: deep learning integration, live updates, and open collaboration
Learning Resources
Access our curated collection of learning materials, tools, and community resources
Our official Instagram page
GitHub Organization
Access our open-source projects
Newsletter
In Workshop π οΈ