SYSTEMS · AI · ENGINEERING
I'm Bobo.
I care about how complex systems remain reliable and efficient in the real world, and how the same engineering mindset can help us build better AI systems.
For more than five years, I worked on the core engineering and technical leadership of real-time collaborative documents at Kingsoft Office. My work spanned system and application architecture, core implementation, code review, performance engineering, and team delivery. I care about more than making a feature run: I want systems to remain understandable, reliable, and evolvable under high concurrency, long-lived connections, and failure.
Since September 2025, I have been pursuing an MSc in Computer Science at Leiden University, with a focus on Advanced Computing and Systems and Artificial Intelligence. This has given me a research lens for ideas I first encountered in production: distributed systems and data-intensive applications, now extended to AI systems, agents, and model inference.
I enjoy taking complex technology apart through its data flow, concurrency model, network protocol, and performance limits, until I can explain, test, and implement it. This blog documents that process through notes on Go and systems programming, distributed systems, Transformers, AI agents, courses, and engineering lessons.
Selected work
REAL-TIME COLLABORATION
Collaborative documents
Worked on architecture, core systems, and backend performance, and received two invention patents covering collaborative-session scheduling and text-aware image compression.
HIGH-PERFORMANCE NETWORKING
Go RPC at connection scale
Designed an event-driven network library with lock-free buffers and zero-copy techniques. At 300,000 long-lived connections, memory usage fell by 84%, while both average and P99 latency decreased by about 57%.
DISTRIBUTED DATA
High-performance cloud K-V database
Built a Go and C++ K-V system with transactions, optimistic locking, consistent hashing, TCC, TTL, and multi-version management, while continuously improving its concurrent access path.
Current interests
I am most interested in systems for AI: reliable runtimes for models and agents, efficient inference and retrieval, and evaluation and feedback loops that enable systems to improve. I also continue to explore distributed data systems, high-performance networking, Go, Rust, and cloud-native infrastructure.
I believe good engineering is not about accumulating complexity. It is about finding the right abstractions so that a system remains understandable as it grows. Writing is one way I turn fuzzy experience into knowledge I can examine and share.
Experience
Leiden University
Computer Science MSc · Advanced Computing and Systems / Artificial Intelligence
Kingsoft Office
Real-time collaborative documents, distributed systems, and high-performance backend engineering, spanning architecture, core development, optimization, and technical leadership.
Software Engineering
Started in software engineering and gradually moved deeper into concurrency, storage, networking, and large-scale data processing.
If you are also working where systems and AI meet, you can find me on GitHub.