AI & Energy Researcher

Building intelligent
systems for clean energy

I research machine learning applications for power systems and distributed energy resources — from load forecasting and DER dispatch optimization to graph neural networks for smart grids.

View Projects Get in Touch
10+
Years in Power Systems
3
Active AI Projects
100h
AI Learning Roadmap
DER
Optimization Focus
About
Where AI meets the grid
I combine deep power systems engineering experience with modern machine learning to solve real-world energy challenges.

My work sits at the intersection of power systems engineering and artificial intelligence. I focus on applying ML techniques — from gradient-boosted trees to transformer models and reinforcement learning — to problems that matter in the energy transition.

I'm particularly interested in DER coordination, microgrid optimization, and using graph neural networks to model complex grid topologies. I believe AI will be a key enabler of a reliable, decarbonized power system.

Outside of research, I document my learning journey in public — including a 100-hour AI roadmap and study guides designed for engineers entering the AI space.

Power Systems

DER Optimization Microgrid Control Load Forecasting Grid Stability Energy Storage SCADA / EMS

AI & Machine Learning

Deep Learning Transformers Reinforcement Learning Graph Neural Networks Time-Series ML LLMs

Tools

PythonPyTorch Scikit-learnPandas CityLearnGrid2OpPyPSA
Projects
Selected work
Research and engineering projects applying AI to power systems challenges.
⚡
DER Dispatch RL Agent
A reinforcement learning agent (PPO) that optimizes battery and solar dispatch in a multi-building microgrid to minimize cost and peak demand. Built on the CityLearn environment with a custom reward function incorporating carbon intensity.
Reinforcement Learning PyTorch CityLearn DER
View on GitHub →
📈
Solar & Load Forecasting
An ensemble forecasting pipeline combining Temporal Fusion Transformers and XGBoost for 24-hour ahead PV generation and load prediction. Evaluated on GEFCom2014 data with MAPE below 3.2% for load and 6.1% for solar.
Transformers XGBoost Time-Series Forecasting
View on GitHub →
🔬
GNN for Fault Location
A Graph Neural Network trained on the IEEE 14-bus test case to localize faults from PMU measurements. Achieves 94% accuracy on unseen fault types without requiring full observability of the network.
Graph Neural Networks PyG Power Flow IEEE 14-bus
View on GitHub →
Learning
Currently studying
I learn in public and document my journey. Here's what I'm working through right now.
🤖
100-Hour AI Roadmap
Deep AI understanding for engineers
  • Mathematical foundations (linear algebra, calculus, probability)
  • Classical ML — supervised, unsupervised, ensembles
  • Deep learning — CNNs, LSTMs, attention mechanisms
  • Transformers & LLMs from scratch
  • AI applications for power systems & DER
Progress
40%
📚
Current Reading
Books & papers on my desk
  • Deep Learning — Goodfellow, Bengio & Courville
  • RL: An Introduction — Sutton & Barto
  • NLP with Transformers — Tunstall et al.
  • Graph Neural Networks in Action — Stamile et al.
  • IEEE Trans. on Power Systems (latest issue)
Books read
3/5
💡
Mastering Claude
10-hour AI assistant roadmap
  • Prompt engineering & chain-of-thought reasoning
  • Building AI workflows for power systems analysis
  • Extended thinking for complex optimization problems
  • Automating research summarization pipelines
Progress
70%
🏗️
Building in Public
Side projects & experiments
  • Personal website — deployed on GitHub Pages
  • DER RL agent on CityLearn environment
  • AI learning roadmap documents & guides
  • Weekly notes on AI × energy systems
Projects live
2/4
Contact
Let's connect

I'm always happy to talk about AI for energy systems, grid optimization, or machine learning in general. Whether it's a research collaboration, a project idea, or just a question — feel free to reach out.


📍 Based in China  ·  Open to remote collaboration worldwide

✉️
Email
twxjtu@gmail.com
→
GitHub
github.com/sinceretw
→
LinkedIn
linkedin.com/in/davidtian
→
🌐
Website
www.wtian.net
→