Skip to main content
HashnodeJiajun's Cyber ​​GardenJiajun's Cyber ​​Garden

Command Palette

Search for a command to run...

Aboutme

Jiajun Wang

M.Sc. Student at FAU | AI Researcher (Generative AI & Medical Imaging)

📍 Erlangen, Germany
📧 [My Email Address]
🔗 [LinkedIn Profile URL] | [GitHub Profile URL]

🚀 About Me

I am currently a Master's student at Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), combining a solid engineering foundation with cutting-edge AI research.

  • Engineering Background: Leveraging experience in telecommunications and industrial automation to solve complex system challenges.
  • AI Research: Published author (BVM Conference) on fine-tuning Vision-Language Models (VLM). Currently researching Diffusion Models for medical image reconstruction.
  • Goal: I am actively seeking an internship in AI, Computer Vision, or Deep Learning. I bring a cheerful personality and a strong ability to bridge engineering constraints with advanced AI solutions.

🎓 Education

Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU) M.Sc. in Electromobility (Focus on AI & Medical Imaging) Oct 2022 – Present (Expected Graduation: Aug 2026)

  • Focus: Deep Learning, Computer Vision, Generative AI.
  • Master's Thesis: Investigating Diffusion Models for medical image denoising, focusing on enhancing generation fidelity and visual perceptual quality.

🔬 Research Experience

Graduate Research Assistant Pattern Recognition Lab, FAU Erlangen-Nürnberg [Date Start] – Present

  • [cite_start]Explainable VLM: Fine-tuned MedGemma-4B using QLoRA for CT image quality assessment[cite: 7, 8]. [cite_start]The model generates radiologist-aligned reasoning, achieving an SRCC of 0.795, significantly outperforming zero-shot baselines (Gemini 2.5 Pro)[cite: 9, 116, 117].
  • [cite_start]Data Pipeline: Constructed a multimodal instruction-tuning dataset by leveraging Gemini 2.5 Pro to synthesize expert-aligned textual explanations[cite: 68].
  • Publication: First-author paper accepted at the German Conference on Medical Image Computing (BVM).

💼 Work Experience

TETRA System Test Engineer Hytera Communications [Date Start] – [Date End]

  • System Deployment: Configured TETRA digital trunking systems for public safety networks, managing physical hardware (BSCU, CHU, Radios) and virtual environments.
  • Testing & Analysis: Executed root cause analysis using Linux, Wireshark, and Xshell, ensuring system stability through rigorous log debugging.
  • Collaboration: Partnered with the German branch (HMF Smart Solutions) to conduct remote joint debugging and resolve critical technical issues using Jira.

Engineering Intern Esquel Group [Date Start] – [Date End]

  • System Design: Designed a PLC-based monitoring system using position sensors to track roller displacement, reducing manual errors and improving yarn quality.
  • Outcome: Secured 1 Invention Patent and 1 Utility Model Patent based on the novel sizing machine startup/shutdown mechanism.

📝 Publications & Patents

Publications

  • Explainable Radiologist-Aligned VLM for CT Image Quality Assessment Accepted at the German Conference on Medical Image Computing (BVM). Proposed a parameter-efficient SFT framework for MedGemma-4B-IT using QLoRA.

Patents

  • Invention Patent: [Patent Name/Number] (Sizing Machine Monitoring System)
  • Utility Model Patent: [Patent Name/Number]

🛠 Skills

  • AI & Deep Learning: PyTorch, LLMs/VLMs (Fine-tuning, RAG), Diffusion Models, PEFT (LoRA/QLoRA), Computer Vision.
  • Programming: Python, C++, MATLAB.
  • Tools & Platforms: Linux, Docker, Git, Wireshark, Jira.
  • Languages: English (Professional), German (Intermediate), Chinese (Native).