🌌 I'm Jiahao Zheng · aka "++"

Undergraduate Researcher at Nanjing University of Posts and Telecommunications (NJUPT)
📍 Majoring in Internet of Things Engineering

I am an AI researcher trying to connect 3D visual understanding, physical-world signals, and interactive generation. My work currently focuses on:

• 🧊 3D Reconstruction & 3DGS - Building stable 3D representations from sparse and imperfect observations
• ⚛️ AI for Quantum Sensing - Learning from noisy physical signals with physics-guided modeling
• 🤖 Robot Vision with Hardware Systems - Connecting perception, deployment, embedded hardware, and action
• 📐 CAD / 3D Content Generation - Exploring structured 3D assets, editable geometry, and design-oriented generation
• 🌐 World-model-oriented Thinking - Moving from "seeing the scene" toward predicting, editing, and interacting with it

💡 It's a clever pun on my Chinese name:

• 🔤 Jiahao sounds like "Plus Sign" (加号)
• ➕ Zheng means "Positive" (正)

So mathematically, I am literally ++ 😄
Like the increment operator - always improving, always growing!

🧊 3D Reconstruction & 3DGS Current focus: Sparse-view 3D Gaussian Splatting Geometry-aware regularization Depth priors & foreground masks Multi-view consistency under imperfect observations Why I care: Good 3D should be more than pretty rendering. It should support understanding, editing, simulation, and downstream tasks.	⚛️ AI for Quantum Sensing Current focus: Diamond NV-center sensing Low-SNR signal reconstruction Physics-guided loss design Robust thermometry and cross-sample generalization Why I care: Noisy physical signals are annoying, but they are excellent teachers of robustness.
🤖 Robot Vision + Hardware Current focus: Real-time perception Vision-guided robotic systems Embedded deployment and hardware-aware AI Perception–decision–action loops Why I care: A vision model should not only recognize things. Sometimes it should help a robot actually do things.	📐 CAD / 3D Generation Current focus: CAD-like structured generation 3D assets and editable geometry AI-assisted design and reconstruction From visual content to usable 3D representations Why I care: The fun part is not just generating pixels, but building worlds that can be edited, simulated, and reused.

I am actively expanding my research horizons into the following frontier areas:

🤖 AI & Deep Learning Frameworks

🦾 Robotics & Embodied AI

💻 Development & Research Tools

🔬 Scientific Computing

🎨 Productivity & Documentation

🎮 Mobile & Others

My engineering practice spans 3D reconstruction, AI4Quantum, robot vision with hardware, and CAD / 3D generation. I enjoy turning loose ideas into runnable systems, visible demos, and occasionally surprisingly stubborn bugs.

🐍 GitHub Contribution Snake

I am an avid marathon runner, believing that the endurance required on the track mirrors the persistence needed in scientific research.

"Every kilometer on the track is a metaphor for every epoch in training neural networks."

What Running Teaches Me:

• Patience - Models converge slowly, just like endurance builds gradually
• Resilience - Hitting the wall at km 30 ≈ NaN loss at epoch 87
• Optimization - Finding the right pace ≈ Finding the right learning rate

My idols, LeBron James and Kylian Mbappé, have profoundly shaped my understanding of dedication and peak performance.

LeBron James The NBA's all-time leading scorer and a defining figure of athletic longevity and leadership.

Kylian Mbappé FIFA World Cup winner and one of the most explosive and talented forwards in modern football.

💭 Random Dev Quote

⭐️ From jiahaozheng406 · "++" · The Incremental Optimist

_{"Coding for AI, Coding for Science. Exploring the Physics of Intelligence." 🌌}