Ashley Li

About me

I'm a Software Engineer from Taiwan, passionate about Computer Vision , Machine Learning , Signal Processing , and Visual related Technologies. With solid experience in machine learning, software systems development, and full-stack web applications .

I recently graduated from The Hong Kong University of Science and Technology with a BEng in Computer Engineering and a minor in Big Data Technology .

Currently, I work as a Research Assistant at Academia Sinica in Institution of Statistics, where I contribute to cutting-edge research projects in Smart Health Project and AI-driven Analysis . My recent research interest lies on Medical AI and Agentic AI , including computer vision related applications and signal processing methods.

Feel free to connect and get in touch with me :)

What i'm doing

Computer Vision

Developing real-time object detection and tracking systems with YOLO, related adaptation areas including Traffic Analysis and People Counting; Researching Agentic AI for advanced medical image analysis.
Machine Learning

Investigating state-of-the-art Agent AI, LLM and Deep Learning technology, including transformer-based and multi-media models for practical applications in Medical related AI and Autonomous Systems.
System Development

Building efficient and scalable software systems using Python and JavaScript, with a focus on performance optimization, data transmission, seamless integration, and deployment.
Web Development

Building full-stack web applications using modern frameworks like React, Node.js, FastAPI, and Redis database storage for seamless user experiences.

Soft Skills

Fast Learner

I have the ability to quickly acquire new knowledge and adapt to changing technological environments.
Teamwork

I collaborate effectively in multidisciplinary projects and communicate clearly with team members, also experienced in leading teams.
Problem Solving

I am creative and efficient in painpointing the issue and solving technical problems.
Effective Communication

I have the skill to convey ideas clearly and concisely, both technically and non-technically.

Technical Skills

Programming Languages

Machine Learning & Data Science

Big Data & Cloud Computing

Web Development & Databases

Parallel & Systems Programming

DevOps & Deployment Tools

Design & Media

Languages

Mandarin
100%
English
80%
Cantonese
30%
Korean
30%

Certificates

Resume

Education

The Hong Kong University of Science and Technology (Hong Kong)
08/2021 - 07/2025

BEng in Computer Engineering with a minor in Big Data Technology.

Graduated with Second Class Honors, Division I (Top 22%, Ranked 25th / 113)

Senior Year GPA: 3.66/4.3 | Cumulative GPA: 3.3/4.3
📚 Coursework
- Machine Learning
- Deep Learning in Computer Vision
- Natural Language Processing
- Parallel Programming
- Computer Architecture
- Operating Systems
- Cloud Computing & Big Data Systems
- Signal and System
- Embedded System
- Object-Oriented Programming
- Probability and Random Processing
- Discrete Mathematics
- Matrix Algebra and Applications
- Calculus I & II, Multivariable Calculus
Chalmers University of Technology (Sweden)
08/2023 - 01/2024

School Nominated Exchange Study in Master Level Courses

Focusing on advanced computer science courses and cultural experiences
📚 Coursework
- Applied Signal Processing
- Computer Graphic
- Image Processing
- Algorithm
National Experimental High School at Hsinchu Science Park (Taiwan)
08/2015 - 06/2021

Graduated with GPA 4.0/4.3

Financial manager of Dance society

Lead Coordinator of Joint-Club Annual Performance

Nationawide Short Essay Competition - Hsinchu Division - A Place

Experience

Research Assistant @ Academia Sinica, Institute of Statistical Science (Taipei, Taiwan)
07/2025 - Present
- Project PI: Prof. Hsin-Chou Yang
- Participated in research projects under Smart Health and Precision Medicine , AI-driven Analysis and support AI-related coding tasks, details can be found at Portfolio section.
⚙️ Skills: Machine Learning, Medical Imaging, Medical AI, Computer Vision, System Development, Web Development
Lead Project Coordinator | FullStack Developer @ HKUST Path Advisor (Hong Kong)
08/2024 - 08/2025
- Project PI: Prof. Raymond Wong
- Progressed from frontend developer , to frontend team leader , finally to project coordinator ; leading 10+ member cross-functional team to deliver new features and manage development cycles.
- Collaborated with campus departments (ITSC & Campus Management Office) to integrate campus map updates and resolve urgent issues, ensuring accurate map synchronization and seamless user experience.
- Optimized internal communication and workflows, establishing documentation and knowledge transfer practices to shorten feature delivery time and improve team efficiency.
- Resolved 20+ user-reported issues and introduced UX enhancements , contributing to a 15% increase in active student users.
⚙️ Skills: Leadership, Communication, Project Management, Full Stack Development
Simulation Team Software Engineer Intern @ ITRI, Autonomous Vehicle Lab (Hsinchu, Taiwan)
06/2024 - 08/2024
- Built a simulation scenario management platform (infrastructure + website) for Autonomous Vehicle control testing, using FastAPI to retrieve and process OpenSCENARIO data and runtime parameters from MongoDB ; streamlining scenario setup, execution, monitoring, and improving data accessibility.
- Automated AV scenario testing with esmini and delivered schematic images and video for visualization of testing results; cutting manual validation effort, reducing in-person testing risks, accelerating iteration speed, and providing intuitive result analysis.
- Integrated webhook-based synchronization for correlated parameters and simulated outputs, eliminating data mismatches and enhancing user experience with a unified overview.
- Achieved 95%+ unit test coverage with Pytest and comprehensive documentation with JSDoc , enabling early defect detection and boosting developer efficiency by making the codebase easier to understand, maintain, and extend.
⚙️ Skills: Python, TypeScript, Autonomous Vehicle, API Development, Webhook Integration, Testing and Documentation, CI/CD, Git
Software Engineer Intern @ NovBee Limited (Hong Kong)
06/2023 - 08/2023
- Developed and deployed 30+ reusable React components with TypeScript , improving code maintainability and accelerating feature development.
- Implemented modular page structures and layouts, reducing frontend code duplication by 40% and ensuring long-term scalability.
- Collaborated with the UI/UX team to enforce design consistency, ensuring 95% alignment with style guidelines and improving usability scores in student feedback surveys.
⚙️ Skills: React, CSS, TypeScript, Collaboration and Communication
Frontend Developer | Frontend Team Leader @ HKUST Path Advisor
09/2022 - 07/2024
- Project PI: Prof. Raymond Wong
- Maintain and improve frontend code using JavaScript for an intuitive and user-friendly web interface.
- Collaborate closely with the ITSC department to ensure seamless integration of features.
- Edit and update campus maps using data received from Campus Management Office, ensuring accuracy and relevancy.
⚙️ Skills: JavaScript, Frontend Development
Software Engineer Intern @ Wiadvance Technology Co. (Taiwan)
07/2022 - 08/2022
- Researched and optimized customer service workflows for a new cloud app release, reducing average user query resolution time by 20% and boosting customer satisfaction ratings.
- Debugged and enhanced backend APIs for the internal website, increasing system reliability by ensuring data fetching correctness.
⚙️ Skills: Python, API Development
UIUX Designer @ USThing
09/2022 - 01/2024
- Designed a team introduction page, introducing team responsibility to the public.
- Created the USTools admin interface and login page, streamlining administrative workflows for future features.
⚙️ Skills: UI/UX Design, User Research, Prototyping
Publicity Manager @ HKUST Taiwanese Students' Association
06/2022 - 07/2023
- Managed official cross-platform social media accounts, growing Instagram followers by 60% (Experience camp account) and 30% (Society main account) within one year, and reaching an audience of 1,000+ students.
- Designed and Created Content for all social media platforms, driving a 40% increase in average post engagement.
- Designed marketing campaigns and event souvenirs that attracted 80+ participants across cultural events, strengthening the Taiwanese student community.
- Authored the society's first branding guideline , providing a long-term framework that subsequent committees have adopted.
⚙️ Skills: Public Relations, Event Promotion, Content Creation
Marketing Team Member @ HKUST Be Our Guest
09/2021 - 06/2023
- Organized 1 online and 3 in-person cultural exchange events at HKUST, engaging over 500 participants.
- Designed compelling visual content using Adobe Illustrator and Final Cut Pro, consistently reaching over 700 accounts.
⚙️ Skills: Marketing, Content Creation, Social Media Management

Awards & Honors

Best Computer Engineering Final Year Project - Champion
2024-2025
Awarded for outstanding Final Year Project in Computer Engineering Major.
HKSAR Government Scholarship - Reaching Out Award
2023
HKUST Undergraduate Exchanged Scholarship.

Project

Quick Draw Duel!

📋 Description

An AI-powered web game that challenges players to outdraw an AI opponent by sketching objects as quickly and accurately as possible, designed for 2025 Academia Sinica Open House event. The game leverage WebSocket, real-time pretrained CNN inference, a FastAPI + Redis backend, QRcode generate certificate delivery, education purposed feature map using UMAP for visualizations, and a live update Streamlit dashboard. Provide an engaging and interactive experience.

⚙️ Key Technology & Contributions

Engineered a responsive multiplayer drawing game frontend with WebSocket integration, enabling real-time sketch input, CNN-based inference, live predicted probability updates, and instant certificate generation across multiple players.
Developed a scalable Python FastAPI backend with Redis, supporting high-throughput real-time data transmission and storage/retrieval; Implemented QR code generation for on-demand certificate downloads and integrated UMAP dimensionality reduction to visualize player drawing clusters against training datasets.
Created an interactive Streamlit dashboard with live statistical data visualization, dynamic player rankings, and real-time performance analytics, enhancing audience engagement during public screening.

🏷️ Keywords

Python, JavaScript, TensorFlow, Keras, CNN Inference, Redis, FastAPI, Streamlit, WebSocket, QRCode, Deployment, Realtime, UMAP dimentionality reduction

People Tracking and Counting for Statistical Analys

Developing an AI System for Smart Traffic: From Vehicle Detection to Data Analytics Dashboard

📋 Description

Final Year Project focused on developing an end-to-end AI solution for real-time traffic monitoring and analytics. The system integrates deep learning–based vehicle detection with multi-object tracking, and an interactive analytics dashboard to deliver actionable insights for urban traffic management in Hong Kong. In collaboration with K11 MUSEA, traffic cameras along Salisbury Road (eastbound and westbound) were analyzed to study flow patterns, providing mall management with data-driven insights to optimize vehicle routing and improve traffic efficiency around the mall.

⚙️ Key Technology & Contributions

Fine-tuned YOLOv11 with public available dataset and self-labeled dataset for robust vehicle detection under diverse traffic and weather conditions, improving detection performance in complex Hong Kong road scenarios.
Integrated ByteTrack for precise multi-object tracking, enabling extraction of key traffic metrics (e.g. vehicle count, speed, flowrate, ...) with high tracking consistency.
Developed a real-time dashboard with Streamlit, powered by FastAPI and Redis Database, to visualize live traffic data and trends for stakeholders.

🏷️ Keywords

Python, PyTorch, YOLO11, ByteTrack, Redis, FastAPI, Streamlit, Computer Vision, Real-Time Analytics, Smart Traffic

Deepfake image classification

📋 Description

Research project in the Deep Learning in Computer Vision (DLCV) course focused on distinguishing real images from AI-generated ones. The study benchmarked multiple CNN and Transformer-based architectures on the CiFAKE dataset and investigated image forensics techniques -- ELA to evaluate their effectiveness in detecting synthetic content.

⚙️ Key Technology & Contributions

Benchmarked CNN architectures (VGG16, ResNet50, MobileNet, EfficientNet), achieving up to 98% accuracy, with EfficientNet identified as the most effective.
Evaluated Vision Transformers (ViT), achieving the highest recall (0.9917) compared to previous CNN-based models.
Explored Error Level Analysis (ELA) with lightweight models, observing a 13% accuracy drop, highlighting the trade-off between model size and image preprocessing technique's effect on detection capability.

🏷️ Keywords

Python, PyTorch, CiFAKE dataset, CNN (VGG16, ResNet50, MobileNet, EfficientNet), Vision Transformer (ViT), Error Level Analysis (ELA), Deepfake Detection, Image Processing

Hong Kong Horse Racing Prediction

📋 Description

Project in the Machine Learning course focused on predicting horse race outcomes as a pointwise Learning-to-Rank problem, with emphasis on the top four finishing positions to align with common betting strategies.

⚙️ Key Technology & Contributions

Conducted race-level features engineering (jockey/horse performance, track condition, race distance), improving the model’s ability to capture performance factors and enhance ranking precision.
Implemented supervised regression models (Linear Regression, MLP, Random Forest), with the MLP achieving the highest top-4 prediction accuracy.
Applied a Learning-to-Rank framework to transform MLP outputs into ordered race predictions, improving ranking consistency and enabling practical betting strategy simulations.

🏷️ Keywords

Python, Pandas, NumPy, Scikit-learn, TensorFlow, Linear Regression, Random Forest, Multilayer Perceptron (MLP), Learning-to-Rank (LTR), Feature Engineering

Image Classification with Data Augmentation via AdaIN Style Transfer

Serverless ML Deployment with OpenFaaS

📋 Description

Project in the Cloud Computing (COMP4651) course focused on deploying a lightweight MobileNet model on Kubernetes to compare serverless (OpenFaaS) and traditional server-based architectures for ML inference, analyzing trade-offs in scalability, latency, and cost.

⚙️ Key Technology & Contributions

Configured a single-node Kubernetes cluster with Minikube and deployed MobileNet via OpenFaaS, achieving real-time inference with ~200ms latency.
Developed and containerized a Python Flask function to serve classification requests, enabling scalable, on-demand inference with dynamic resource allocation.
Benchmarked serverless vs. server-based deployments, demonstrating lower maintenance overhead and identifying trade-offs in scalability, cost, and latency.

🏷️ Keywords

Python, Flask, OpenFaaS, Kubernetes, Minikube, Docker, MobileNet, Serverless Computing, Cloud Deployment, ML Inference

MATLAB Portraiture Image Processing

📋 Description

Project in the Image Processing course focused on developing a multi-stage pipeline in MATLAB to simulate “portrait mode,” enhancing subject clarity while applying background blur for a professional portrait effect.

⚙️ Key Technology & Contributions

Implemented preprocessing techniques (histogram equalization, sharpening, lowpass filtering) to improve subject contrast and reduce noise, yielding sharper facial details across all test images.
Evaluated multiple edge detection methods (Laplacian filter, wavelet decomposition, binary thresholding), with binary thresholding delivering the highest segmentation accuracy and lowest noise in both simple and textured backgrounds.
Applied black–white mask manipulation (imclose, imfill, bwareaopen) to remove noise clusters and reconstruct clean subject masks, producing portrait images with consistent subject isolation and realistic bokeh effects.

🏷️ Keywords

MATLAB, Image Processing, Histogram Equalization, Sharpening, Lowpass Filtering, Laplacian Edge Detection, Wavelet Decomposition, Binary Thresholding, Mask Manipulation, Background Blur, Portrait Mode

Terrain Rendering with OpenGL

📋 Description

Project in the Computer Graphics course focused on developing a terrain rendering system using heightfield textures and procedural height maps to achieve realistic visualization with dynamic lighting and shading.

⚙️ Key Technology & Contributions

Generated and rendered a triangle mesh in a Vertex Array Object (VAO) with dynamic vertex displacement for real-time terrain deformation.
Implemented vertex shader operations to sample heightfields, compute offsets, and derive normals, ensuring accurate lighting and shading across varied terrain features.
Applied texture mapping with aerial photos, achieving enhanced visual fidelity and realism in terrain surfaces.

🏷️ Keywords

C++, OpenGL, GLSL, Vertex Shader, Heightfield Texture, Procedural Terrain, Mesh Generation, Texture Mapping, Real-Time Rendering

Arduino Line-Following Car

📋 Description

Project in the Introduction to Electro-Robot Design course focused on designing and implementing an autonomous car capable of following a white line track and responding to logic-dependent conditions such as turn and obstacles.

⚙️ Key Technology & Contributions

Designed and integrated dual optical sensors and a bumper sensor with Arduino logic, enabling the car to detect track lines and obstacles.
Optimized motor control through PWM tuning and H-bridge circuitry, balancing speed and precision to successfully navigate curves, 90° turns, and forked paths.
Implemented real-time decision logic in Arduino C to handle branching paths, obstacle detection, and corrective movements, achieving successful runs during lab testing despite hardware failures.

🏷️ Keywords

Arduino, C Programming, PWM, H-Bridge, Optical Sensors, Bumper Sensor, Logic Gates, Embedded Systems, Autonomous Vehicles, Robotics

What i'm doing

Computer Vision

Machine Learning

System Development

Web Development

Soft Skills

Fast Learner

Teamwork

Problem Solving

Effective Communication

Technical Skills

Programming Languages

Machine Learning & Data Science

Big Data & Cloud Computing

Web Development & Databases

Parallel & Systems Programming

DevOps & Deployment Tools

Design & Media

Languages

Mandarin

English

Cantonese

Korean

Certificates

Hugging Face LLM

Hugging Face AI Agent

Statistics Course

Algorithm Course

Optimization for DS and ML

Education

The Hong Kong University of Science and Technology (Hong Kong)

📚 Coursework

Chalmers University of Technology (Sweden)

📚 Coursework

National Experimental High School at Hsinchu Science Park (Taiwan)

Experience

Research Assistant @ Academia Sinica, Institute of Statistical Science (Taipei, Taiwan)

Lead Project Coordinator | FullStack Developer @ HKUST Path Advisor (Hong Kong)

Simulation Team Software Engineer Intern @ ITRI, Autonomous Vehicle Lab (Hsinchu, Taiwan)

Software Engineer Intern @ NovBee Limited (Hong Kong)

Frontend Developer | Frontend Team Leader @ HKUST Path Advisor

Software Engineer Intern @ Wiadvance Technology Co. (Taiwan)

UIUX Designer @ USThing

Publicity Manager @ HKUST Taiwanese Students' Association

Marketing Team Member @ HKUST Be Our Guest