Transferred Correlated Learning
The earlier era of neural networks, where deep learning didn't exist. This includes a hand-crafted 3-layer perceptron combined with transfer learning.
Neural NetworksTransfer LearningResearch
I'm a software engineer and founder.
I build domain-intelligent applications—tailored AI solutions designed for specific, deep-rooted domain-specific problems that one-size-fits-all models simply may not solve.
Crafting thoughtful knowledge structure, data pipelines, and data-savvy applications powered by AI.
About
Background
Currently employed by a big tech, my professional journey as a software engineer has been around every aspect of data and the infrastructure that enables it: data pipelines, search engine, knowledge graph, cloud databases and observability (check my Linkedin on the contact section).
I believe that AI can and will penetrate into many fields and should do it the right way. Workflows can be generalized and optimized, but knowledge and data, often with irregularity and noise, are the real treasure that can unlock the potential of AI.
When I'm not working on projects, you may find me playing with my daughter, watching sports or listening to my vinyl record collection.
Selected Projects (personal/academic only)
Causal Discovery on international relations from GDELT
Finding out the underlying causal network among international events using the GDELT dataset. This project involves time series data, structured causal models, and some natural language processing
Causal DiscoveryTime SeriesNLPResearch