Welcome to my GitHub page! 
 My name is Titus and I'm a data scientist from  Singapore     
🏫 I’ve a Masters degree in Artificial Intelligence 👨🎓 but I'm forever playing catch-up with the rapidly developing field of data science 🤣. Nonetheless, those nights spent learning the fundamentals of AI (computer vision, natural language processing, recommender systems, machine learning, machine learning engineering, etc) were very memorable!
🤖  I genuinely enjoy building LLM applications (agents, agentic RAG systems, etc.) and basically live on the Llamaindex / Langchain / Chainlit / Autogen / CrewAI's documentation pages!
✍️  I author and publish medium articles on generative AI frameworks for fun! I've published 40 articles so far! 
🦾  I'm also a citizen RPA developer. UiPath is awesome! 
 
🌱 I’m hoping to learn more about:
- 
    Java - because almost every company uses Java ♨️!
 
 
- 
    Full stack development to become a more complete developer - coding is so fun 👨💻!
 
 
- 
    Data engineering - Developing robust data pipelines 🛠️!
 
   
- 
    Quantum Computing - I believe it's the future and it's coming sooner than we think! ⚛
 
 
| 🚀 Check out my personal webpage! | 
| 🎁 Projects | 📝 Summary | ⭐ Domains | 🛠️ Tools | 
| The Margin Call | LLMs for investing - technical analysis, fundamental analysis, AutoRAG, online research and AutoML to forecast stock prices |  |     | 
| A Gentle Introduction to the LLM Multiverse | The GitHub repository used for all Medium articles introducing LLM frameworks |  |     | 
| You Only Edit Once | Using object detection models to trim and beautify diving videos - feel free to try it out! |  | |
| A Song for Singapore | Using LSTM, GRU, Seq2Seq models and GPT-2 to write lyrics, melody, harmony and sing out Singapore's next national day song! |     | |
| The Coffee Recommender System | Using deep learning and collaborative filtering to develop Singapore's coffee recommendation system. |  | |
| AI Backdoor Catcher | Using neural cleansing, neural attention distillation and other advanced methods to identify backdoors embedded in AI models | ||
| Face Mask Classification | This was an attempt at developing and quantizing image classification models for deployment to edge device through GCP to support the government's COVID-19 face mask detection enforcement measures. |  | |
| Skimlit | The aim of this project is to make chunky abstracts/summaries appear skimmable,and follows the experiments (and data) by the paper: [PubMed 200k RCT: a Dataset for Sequenctial Sentence Classification in Medical Abstracts](https://arxiv.org/abs/1710.06071). |  | |
| FoodVision-Tensorflow | This project is a personal project to develop a computer vision classification model for food images. |  | |
| Improving NYC 311 Efficiency | Forecasting closure duration of 311 complaints to support NYC agencies in case prioritization |  |  | 
| The IoT Liquid Pipeline Monitoring System | Developing an entire proof-of-concept for IoT-driven pipelines using Arduino, SQL and Python. | 
















