Hi! 👋 I'm Kevin Pek

I'm currently a Software Engineer at Meta. Here you can find some of my random side projects. Occasionally I ramble about random things on my blog.

Resume

About Me

Born and raised in Singapore, I graduated from NUS at the end of 2024. I first learnt programming in 2013 to make video games, and that kickstarted my love for learning random knowledge and skills.

I enjoy singing, art and movies, and I am always down to learn or experience something new.

Skills & Interests

AI/ML

    I started learning about this through the course "Machine Learning A-Z" on Udemy in 2021. Around this time I also came across NVIDIA's GauGAN, which started my fascination with Deep Learning and Generative AI.

    My undergraduate coursework and internships have given me a deeper understanding of this topic, and I also play with various AI models in my side projects too.

    Web Development

      I started my journey into web development in 2021 from Harvard's CS50W course. Since then it has become a skill I use frequently for various side projects.

      Digital Art

        I originally learnt blender and photoshop to make assets for the games I wanted to make. This ended up spinning off into a new hobby.

        Recently I have been learning to make interactive/generative art that mixes computer graphics with machine learning techniques.

        Past Experiences

        Data Scientist Intern @ GovTech

        • Python
        • FastAPI
        • PyTorch
        • PaddlePaddle
        • sklearn
        • Docker
        • Deep Learning
        • AWS
        I made a custom implementation of the Tip-Adapter, a few shot learning technique for CLIP models. I also finetuned multimodal encoder models on Sagemaker, deployed them on a FastAPI server, and made performance optimisations for inference.

        Research Assistant @ NUS IORA, NCS

        • Python
        • Data Analysis
        • FiPy
        • PyTorch
        • Azure
        I did data analysis on time series data and trained forecasting models for predicting temperature and power usage in data centers, and created thermal renders of temperature distribution.

        Research Assistant @ NUS IORA, SIA

        • Typescript
        • Express
        • React
        • Docker
        • Python
        • FastAPI
        • Multi-armed Bandits
        • AWS
        I implemented a research paper about Multi-arm Bandit algorithms for question recommendation into a FastAPI service. I also built a quiz platform with user management dashboards and deployed these services using docker.

        Full Stack Intern @ Idemia

        • Java
        • Spring Boot
        I developed an orchestrator application in Java using Spring Boot and Apache Camel, and did bug fixes for various services.

        Modelling and Simulations Intern @ DSTA

        • C#
        • Unity
        • Reinforcement Learning
        I created a drone flight simulation training environment in Unity. Used Unity's ML Agents package to train reinforcement learning agents to be used alongside computer vision models and octree bsaed 3D pathfinding algorithm.

        Other Things I've Made