Building My Agent Harness
Rationale
Continue readingHi, I’m Xiao. I’m a software engineer interested in search, AI, and distributed systems. This is where I write about what I’m building, learning, and still trying to understand.
A learning journal, in public. More about me ↗
SEPTEMBER2026
MAY2026
Building Peregrine
During a trip back to China, I finally had some quiet time to write down what I want to remember from building Peregrine: a small Zig + Metal inference engine for Qwen3.5 9B q4 on Apple Silicon.
Continue readingJANUARY2026
Exploring RLM Part 2: Context Engineering for Coding Agents
In Part 1, I explored the RLM paradigm - giving LLMs a REPL to programmatically interact with data instead of processing massive contexts in a single forward pass. This post covers how I’ve been applying RLM and other techniques to manag...
Continue readingExploring Recursive Language Models
I’ve been exploring Recursive Language Models (RLM), a new inference paradigm from MIT’s OASYS lab. The core idea is compelling: instead of forcing LLMs to process massive contexts in a single forward pass, give them a Python REPL and le...
Continue readingDECEMBER2025
Titans Part 3: HOPE Architecture - From Paper to Reality (and Back)
This is Part 3 of my Titans implementation series. Part 1 covered the basic memory mechanism. Part 2 focused on performance optimization. This post tackles the full HOPE (Hierarchical Optimized Parallel Encoding) architecture - the multi...
Continue readingTitans Part 2: Optimizing Memory Updates and Adaptive Learning
This is a follow-up to my previous post on implementing Titans. After getting the basic implementation working, I dove deeper into performance optimization with Claude Code as a pair programmer. This post covers the journey from a workin...
Continue readingImplementing Titans - Learning to Memorize at Test Time
Context engineering has become extreme intriguing to me recently, as hands on building couple of agentic platform project, one asepct of context engineering is memory and continous learning, Manus shared an excellent learning: Context En...
Continue readingBuilding SGREP - Part Two
After shipping the first version of sgrep with ColBERT late interaction, I thought the hard work was done. The search accuracy was good - MRR of 0.70 on my test queries, significantly better than plain semantic search. But when I started...
Continue readingBuilding SGREP
Recently, there was a great mrep tool published by mxbread team, helping to address the issue that LLM harnesses such as claude code, codex, amp, when doing search, spent unnecessary time to retrieve useless tokens. Here’s what mgrep claim:
Continue readingNo posts found. Try another word or topic.