8000 JintongHe (Jonathan (Jintong) He) · GitHub
[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Skip to content
View JintongHe's full-sized avatar

Block or report JintongHe

Block user

Prevent this user from interacting with your repositories and sending you notifications. Learn more about blocking users.

You must be logged in to block users.

Please don't include any personal information such as legal names or email addresses. Maximum 100 characters, markdown supported. This note will be visible to only you.
Report abuse

Contact GitHub support about this user’s behavior. Learn more about reporting abuse.

Report abuse
JintongHe/README.md

Jintong (Jonathan) He – Portfolio

📍 MS in AI Engineering | Carnegie Mellon University
📧 jintongh@andrew.cmu.edu
🔗 LinkedIn Profile


About Me

I am a highly motivated AI engineering student seeking internship and full-time opportunities in machine learning engineering and data science. Currently pursuing a Master’s degree in AI Engineering – Mechanical Engineering at Carnegie Mellon University, I bring a strong foundation in machine learning, computer vision, and data analytics.

As a Machine Learning Engineer at Zhongke Xingqi Technology Co., Ltd., I led a team to develop high-precision object detection models for satellite imagery, achieving over 90% precision and recall. I also optimized GPU-based model deployment, significantly reducing inference time through parallelization strategies.

Beyond technical skills, I value mentorship and team collaboration, having trained new engineers in deployment and optimization best practices. I am passionate about delivering robust AI solutions and fostering continuous learning environments.


Resume

🔗 View My Resume


Engineering Projects

Satellite Image Object Detection System

📍 Zhongke Xingqi Technology Co., Ltd. | Aug 2023 – Jun 2024

Objective:
Automate large-scale remote sensing tasks for environmental monitoring and urban planning using deep learning.

Technologies: PyTorch · YOLOv8 · Vision Transformers (ViT) · Flask · GPU Optimization · Multi-threading

Key Contributions:

  • Model Development: Designed and fine-tuned YOLOv8 and ViT-based models, achieving 90% precision and recall on diverse terrains.
  • Deployment Optimization: Built Flask APIs and applied multi-threading and batching, reducing inference latency by 30%.
  • Team Leadership: Directed a 4-member team, applying Agile practices for integration with downstream analytics pipelines.

Model Visualization Deployment System


Defect Detection System for Wire Arc Additive Manufacturing (WAAM)

📍 University of Wollongong | Jul 2022 – Jul 2023

Objective:
Enhance the safety and efficiency of metal 3D printing through real-time defect detection.

Technologies: Python · PyTorch · CNNs · Computer Vision

Key Contributions:

  • Signal-Based Feature Engineering: Extracted voltage and current features, boosting model F1 score by 5%.
  • Model Development: Built a CNN classifier achieving 99.71% F1 score on the test set.
  • Research Communication: Created technical visualizations and delivered a conference-level poster presentation.

Poster Presentation


Real-Time Pedestrian and Cyclist Recognition (PACR)

📍 University of Wollongong | Feb 2022 – Nov 2022

Objective:
Develop a low-cost, real-time system to improve driver awareness and road safety.

Technologies: Jetson Nano · Python · TensorFlow · OpenCV

Key Contributions:

  • Efficient Object Detection: Developed a lightweight MobileNetV2-based model integrated with IR-switching cameras.
  • Embedded System Optimization: Achieved 18 FPS real-time inference on Jetson Nano.
  • Full System Integration: Tested robustness under diverse urban and lighting conditions.

View Demonstration


Data-Driven Advertising Strategy Optimization

📍 Feng Bian Technology Co., Ltd. | Oct 2021 – Jan 2022

Objective:
Refine advertising strategies for beginner Python courses to improve marketing ROI.

Technologies: Pandas · SQL · Seaborn · Matplotlib

Key Contributions:

  • Data Cleaning and Analysis: Processed and analyzed survey data from over 10,000 customers.
  • Insightful Visualization: Created compelling visuals that influenced business strategy across departments.
  • Business Impact: Recommended optimizations leading to an 8% increase in profitability.

Justification Memo

To: Dr. Alan Thomas Kohler
From: Jonathan He
Date: April 25, 2025
Subject: Application of Communication Skills Learned in Class to Other Coursework


Skills Learned

  • Assertion-Evidence Slide Design
  • IMRD (Introduction, Methods, Results, Discussion) Format
  • Effective Use of Hand Gestures
  • Known/New Information Structuring

Application of Skills

1. Assertion-Evidence Slide Design

In course 24-665 (Wearable Health), I applied the assertion-evidence approach to all presentation slides. Instead of using vague titles, each slide began with a clear, declarative statement summarizing the key takeaway. This significantly improved audience comprehension and engagement. As a result, I earned a perfect score (10/10) on the presentation.


2. Effective Use of Hand Gestures

During the Wearable Health presentation, I deliberately used hand gestures to support my verbal communication:

  • Topic Introduction Gesture: Used a broad open-hand motion to introduce new topics.
  • Bullet Emphasis Gesture: Used finger-pointing motions to guide the audience through key bullet points.
  • Summary Sweep Gesture: Used a sweeping motion to summarize major takeaways.

These gestures helped provide visual structure for the audience and emphasized key transitions during the talk.


3. IMRD Format in Report Writing

In course 24-782 (AI Engineering Project), our final report strictly followed the IMRD structure:

  • Introduction: Introduced the background of imitation learning and its relevance to humanoid locomotion.
  • Methods: Described our approach to applying imitation learning algorithms to train a walking agent.
  • Results: Presented experimental outcomes, performance metrics, and visual results.
  • Discussion: Analyzed the strengths, limitations, and proposed future work.

This structure improved readability and made our logical progression clear to readers.


4. Known/New Information Structuring

Throughout the AI Engineering report, I applied the known/new strategy to improve coherence. For example, after explaining what imitation learning is (known), the next section naturally transitioned into how our team applied it (new). This technique made the report smoother and easier to follow.


Appendix

Attached materials:

{8FEF2984-FF63-4E4A-A561-51198A83167A} {6E09E89F-EF9D-4B32-8478-410E2867FDDC}

Popular repositories Loading

  1. YOLOv8-Train-Pipeline YOLOv8-Train-Pipeline Public

    Python 1

  2. road_network_detection road_network_detection Public

    Using semantic segmentation to detect road networks

    Python 1

  3. YOLOv8_deploy_pipeline YOLOv8_deploy_pipeline Public

    Codes on deploying YOLOv8 on V100 GPUs for fast inference on salellite images

    Jupyter Notebook 1

  4. skills-introduction-to-github skills-introduction-to-github Public

    My clone repository

  5. skills-communicate-using-markdown skills-communicate-using-markdown Public

    My clone repository

  6. Test-Toolchains Test-Toolchains Public

    14763 HW1

0