Summary
A versatile professional with a Bachelor's in Computer Science and a Master's in Industrial Design Engineering. My expertise spans coding, data analysis, and problem-solving, enabling me to manage complex datasets and systems effectively. I combine technical acumen with practical experience in product design, 3D modeling, and project management, offering a holistic understanding of product development and international trade. My hands-on projects and internships have honed my ability to integrate technology and design, making me uniquely equipped for roles in foreign trade that demand analytical, technical, and design skills for navigating global markets and product innovation.
Education
Zhejiang University
Masters in Industrial Design Engineering
September 2023 | Zhejiang, China
Yunnan University
Bachelor in Computer Science and Technology
2019-2023 | Kunming, China
Experience
ProFabx
Software Engineer and 3D Designer
June 2024 - September 2024 | Ningbo, China
- Developed advanced 3D models using Autodesk Fusion 360 and Inventor, successfully transitioning from manual design to automated workflows.
- Created robust Python scripts to automate complex 3D modeling processes through the Autodesk Inventor API.
- Designed an intuitive web interface using JavaScript that dynamically updates 3D models based on real-time parameter changes.
- Implemented a sophisticated parametric modeling system, ensuring dimension modifications automatically propagate throughout entire designs.
- Streamlined design processes by 40% through the strategic automation of repetitive modeling tasks, significantly boosting efficiency.
Projects
BepsBot – AI-Powered Mental Health Peer Support Platform
2025 | Zhejiang University
Tools: Python, Flask, FastAPI, PyTorch, RoBERTa, LangChain, DeepSeek LLM, Sentence Transformers, ChromaDB, RAG, Pushshift API, ThreadPoolExecutor
- Developed a fully functional web-based prototype as an intelligent writing assistant for peer supporters, providing real-time feedback on Emotional Support (ES) and Informational Support (IS) levels and using Generative AI to suggest context-aware improvements for safer, more empathetic replies.
- Designed and deployed a scalable microservice architecture, integrating a large language model (LLM), retrieval augmentation generation (RAG), and a Transformer-based regression model.
- Built a robust data pipeline to crawl and preprocess mental health discussions from Reddit over the past six years. Led a rigorous annotation effort with three domain experts to annotate IS/ES metrics, creating a high-quality gold-standard dataset for supervised learning.
- Experimented with SVM, Multinomial Logistic Regression, Random Forest (RF), and XGBoost to predict support metrics, achieving the best results for IS with RF and for ES with XGBoost.
- Fine-tuned the RoBERTa model on the expert-annotated dataset to predict support quality scores highly correlated with human judgment.
- Built a RAG system using LangChain, ChromaDB, and SentenceTransformers, retrieving semantically similar, high-quality human responses to ground the DeepSeek LLM, reducing hallucinations and improving generated suggestions.
- Deployed models via a high-concurrency FastAPI inference service, decoupled from the Flask frontend, optimizing inference latency by ~65% through asynchronous request processing and parallel execution of LLM generation and security checks (using ThreadPoolExecutor).
- Developed a real-time security filtering middleware using LLM to detect and block harmful content, self-harm risks, and dangerous medical suggestions, ensuring strict adherence to ethical guidelines for mental health applications.
- Implemented an automated active learning pipeline that filters, pseudo-labels, and incorporates new user submissions into the training set, enabling the model to adapt and improve over time without human intervention.
Visit BepsBot
SPROUT – Shape-changing Meditation Device for Academic Mindset Support
Oct 2024 – Dec 2024 | Zhejiang University
Tools: Arduino IDE/C++, Embedded Systems, Sensor Integration, Fusion 360
- Designed a bio-inspired, shape-changing interactive board to help students transform academic challenges into opportunities for growth.
- Addressed Fear of Failure by developing a supportive non-intrusive intervention device for university environments.
- Integrated sensors and actuators with Arduino-based embedded systems to create a meditative interaction, fostering mindset reflection and improvement.
- Utilized the metaphor of "planting a seed" to symbolize the transformation of negative emotions into resilience and a growth mindset.
- Key skills: System Design, Embedded Systems, Circuit Prototyping, C++ Programming, Sensor Integration, Prototype Development.
Autodesk Inventor API with Python – Parametric CAD Automation
Apr 2024 – Aug 2024
Tools: Autodesk Inventor, Python, win32com.client
- Developed a robust parametric modeling system using Python to dynamically update Autodesk Inventor 3D models by modifying key dimensions.
- Automated repetitive CAD tasks, significantly reducing design time and minimizing manual errors in complex projects.
- Seamlessly integrated Python scripting with Inventor’s API to streamline mechanical design workflows and enhance prototyping efficiency.
InclusiveVision - Smart Glasses for Visually Impaired Individuals
December 2023
Tools: Arduino UNO, Ultrasonic Sensors, Fusion 360, Flashforge 3D Printing, Talkie Library, AutoCAD, Laser Cutting
- Developed a functional prototype of smart glasses using Arduino UNO and ultrasonic sensors, achieving real-time obstacle detection with high accuracy.
- Implemented an effective voice alert system using the Talkie library, providing immediate auditory feedback based on sensor data.
- Created optimized 3D models of the glasses frame using Fusion 360, prioritizing comfort and functionality for a wearable form factor.
- Designed an integrated component with multiple ultrasonic sensors for easy assembly and seamless compatibility with existing assistive technologies, enhancing user independence and safety.
IoT Based Dormitory Attendance System with Face Recognition
2022 - 2023 | Yunnan University
Tools: ESP32 Cam, Arduino Mega, OpenCV, MongoDB, GSM Module, LBPH Algorithm
- Developed a contactless attendance system using ESP32 Cam with face recognition to replace manual dormitory attendance tracking.
- Implemented LBPH (Local Binary Patterns Histogram) algorithm with OpenCV for accurate face detection and recognition.
- Integrated real-time data storage using MongoDB cloud database with automated SMS alerts to parents via GSM module.
- Designed a comprehensive IoT architecture with perception, network, middleware, and application layers for robust data management.
- Created a web interface for dormitory managers to monitor attendance records and generate reports in real-time.
Product recommendation system based on Collaborative Filtering Algorithm
2022 - 2023
Tools: Python, NumPy, Pandas, MongoDB, Matplotlib, Flask
- Developed a collaborative filtering strategy for the goodbooks-10k dataset to generate high-quality recommendations based on user purchase and rating history.
- Utilized Python with NumPy, Pandas, and Matplotlib for comprehensive data analysis and visualization, leading to accurate and insightful recommendations.
- Created a personalized recommendation model based on matrix factorization, ensuring tailored suggestions for users.
- Applied Euclidean Distance for similarity analysis between books, significantly enhancing the accuracy of book recommendations.
Publications
- Islam, M., Xue, M., Afra, T. (2025). Exploring Psychologist Applied Biomarkers in Bipolar Disorder: A Systematic Framework. In: Yamanishi, R., Formico, A., Sugiura, Y., Matsumura, K. (eds) Entertainment Computing – ICEC 2025 IFIP TC 14 Workshops. ICEC 2025. Lecture Notes in Computer Science, vol 15935. Springer, Cham. https://doi.org/10.1007/978-3-032-02534-0_8
- Afra, T., Xue, M., Islam, M. (2025). Enhancing Biofeedback Interventions for Depression and Anxiety Through Entertainment Computing: A Systematic Review. In: Yamanishi, R., Formico, A., Sugiura, Y., Matsumura, K. (eds) Entertainment Computing – ICEC 2025 IFIP TC 14 Workshops. ICEC 2025. Lecture Notes in Computer Science, vol 15935. Springer, Cham. https://doi.org/10.1007/978-3-032-02534-0_6
Honors & Awards
United Nations Technology Bank
Global Youth Talent Program
Selected Participant | 2023
- Selected among top young designers worldwide for this prestigious UN program, recognizing innovative talent.
- Recognized for developing innovative 3D design and engineering solutions with real-world applicability.
- Actively participated in a specialized curriculum bridging advanced design principles with sustainable development goals.
- The program's impact and my participation were prominently featured on the official UN Technology Bank website.
Skills
Programming Languages:
Python
C
JavaScript
Web Technologies:
HTML5
CSS3
JavaScript
React.js
Node.js
Database Systems:
MongoDB
Data Science & Machine Learning:
TensorFlow
Scikit-learn
Pandas
NumPy
Matplotlib
Computer-Aided Design & Automation:
Autodesk Inventor API
Parametric CAD Modeling
3D Modeling
Design Automation
win32com (COM Automation)
API Integration
Batch Processing
Simulation and Analysis
Product Visualization