Greeting 👋 Check My portfolio website
Discipline is freedom. True independence comes from mastering yourself, not being a slave to momentary desires.
-
MS, Applied Data Science
University of Southern California (GPA: 3.85/4.0)
Aug 2023 – Dec 2025 -
BS, Computer Science (First Class, GPA: 3.7/4.0)
University of Liverpool
Sep 2019 – Jun 2023 -
BS, Information and Computing Science (GPA: 3.7/4.0)
Xi’an Jiaotong–Liverpool University
Sep 2019 – Jun 2023
💻 Machine Learning Engineer Intern
Henan Changlu Group Real Estate Co., Ltd.
Jan 2025 – Apr 2025
- Implemented an AI-driven system for parking management, encompassing occupancy tracking and multi-violation detection.
- Implemented a PyTorch dataset class for KITTI and built the KITTI API by transforming the labels to a COCO-like format.
- Fine-tuned a pre-trained DETR model by customizing its final classification layer to align with KITTI’s label mapping and conducting extensive hyperparameter optimization, resulting in a 12% increase in detection performance (average AP).
- Enabled seamless model evaluation with enhanced performance metrics and reporting.
💻 Research Assistant
USC Mark & Mary Stevens Neuroimaging and Informatics Institute (INI)
Feb 2024 – Jan 2025
- Developed a DCN-GAN in PyTorch to generate cortex thickness meshes for data augmentation and Mesh Diffusion modeling, enhancing dataset diversity and training robustness.
- Designed and implemented the model with ResNet blocks and attention layers to improve depth and feature extraction, reducing FID for mesh generation by 41.52%.
- Optimized and fine-tuned training on GPU clusters using YAML-based configuration control with OmegaConf; implemented mesh visualization using SciPy for enhanced interpretability.
- Improved GAN stability by applying weight clipping for WGAN and gradient penalty (WGAN-GP), achieving balanced discriminator–generator convergence.
💻 Software Engineering Intern
Eth Tech
Nov 2023 – Jan 2024
- Implemented a distributed e-commerce database system using MongoDB with the Amazon Fashion Products dataset; built backend CRUD operations with Tkinter and PyMongo plus a Streamlit-based desktop query interface to enhance user experience.
- Optimized data distribution by designing an effective hash key for MongoDB sharding, reducing query latency by 50% and improving data reliability through distributed backup strategies.
- Built a lightweight real-time database on AWS EC2, leveraging cURL to communicate via a REST API for efficient data storage and retrieval.
- Visualized overall transaction data using D3.js, Vue and Deck.gl, rendering interactive, map-based dashboards to reveal spatial distribution and transaction patterns.
💻 Machine Learning Engineer Intern
Yisen Tech
Aug 2023 – Oct 2023
- Developed a hybrid recommendation system for personalized user recommendations using Yelp rating data.
- Processed large-scale data using Spark RDD, optimizing feature extraction and transformation for efficiency.
- Improved item-based collaborative filtering by integrating an XGBoost-based model, incorporating key features, and validating their impact through RMSE analysis.
- Achieved 95.1% accuracy, representing a 6% performance boost compared to the baseline.
💻 CNN-Based Brain Tumor Segmentation
- Implemented a U-Net CNN model for brain MRI image segmentation, conducted hyperparameter tuning and architectural optimization, achieving a 0.95 dice score for tumor segmentation.
- Earned the Best Computer Vision Project award at Imperial College London's Data Science Summer Camp 2022.
💻 Object Detection with DETR on KITTI Dataset
- Implemented a PyTorch dataset class for KITTI and build the KITTI API by transforming the label to coco-like format.
- Fine-tuned a pre-trained DETR model by customizing its final classification layer to align with KITTI’s label mapping and conducting extensive hyperparameter optimization, resulting in a substantial increase (12%) in detection performance as measured by average AP and enabling seamless model evaluation with enhanced performance metrics.
- Machine Learning & Deep Learning: PyTorch, TensorFlow; image segmentation, object detection, object tracking, generative models, word embeddings, GPU cluster acceleration
- Data Management & Visualization: MongoDB, MapReduce, distributed file systems; data mining, D3.js, Google Charts, Matplotlib, ggplot2
- Frontend & Tools: React.js, Vue, Streamlit, D3.js; proficient in Python, Java, C++, C#, SQL, R
I’m always excited to collaborate on innovative projects and explore advances in data science, machine learning, and software engineering. Feel free to reach out:
- Email: kimigj999666@gmail.com
- Connect on LinkedIn
- Explore more on GitHub
Love & Peace,
Kimi