
Christopher Tan
Information Eng And Media
School of Electrical & Electronic Engineering
Summary
I chose the Engineering route since Polytechnic because I enjoy the hands-on experience of solving problems. More so, I am deeply interested in learning about new technologies, like AI and Machine Learning. Although my programming skills are not the best, my enthusiasm led me to take up two related projects, one of which uses ML frameworks like BERT to generate embeddings for a large dataset of 1000+ Reddit comments, and then uses classification models like Multilayer Perceptron to predict sarcasm in text, achieving a 71% accuracy. It was a challenging task, but it was also an equally fascinating experience, and I hope to learn more about it and other emerging technologies to come.
Featured Skills
Work Experience
Chemical-Mechanical Polishing (CMP) , Intern Sep 2020 - Jan 2021
Micron Singapore SingaporeCollaborated with one senior technician to design and install a Pad-conditioner drop sensor system to 3 CMP machines, automating fault detection, reducing wafer excursion and yield losses, and improved output efficiency
Education
Bachelor Of Engineering (Information Engineering And Media) (Electrical Engineering) Aug 2023 - Present
Nanyang Technological University SingaporeOther Skills
Careers Christopher favourited
Christopher Tan's Skill Profile
Bachelor Of Engineering (Information Engineering And Media) @ Nanyang Technological University
When I was working at Nanyang Technological University (Aug 2023 - Present), I demonstrated the Machine Learning skill when I was in charge of a group project that aimed to predict sarcastic remarks using machine learning frameworks and models. To tackle this challenge, I proactively researched on machine learning models and looked up tutorials on training a model. As a result, I managed to achieve 71% accuracy using the MLP model.