Blog
-
DASHSys 2026 Keynote
I spoke at DASHSys 2026, the VLDB 2026 workshop on systems for data-centric agents with human-in-the-loop, sharing lessons from building QwenPaw and applications on top of it.
-
Two Claudes
Claude Monet painted the same haystack over and over — at dawn, at dusk, in snow, in summer. He started late in the year of 1890 and kept…
-
Please Use Streaming Workload to Benchmark Vector Databases
Why static workload is insufficient and what I learned by comparing HNSWLIB and DiskANN using streaming workload
-
Finding Needles in a Haystack — Search Indexes for Jaccard Similarity
From basic concepts to exact and approximate indexes
-
GPT-4’s Maze Navigation: A Deep Dive into ReAct Agent and LLM’s Thoughts
In this post I dissect a GPT-4 ReAct Agent capable of navigating a maze, and reveal GPT-4’s thoughts and built-in skills.
-
Human-Aligned Text-to-SQL Evaluation
In my last post about Text-to-SQL using GPT-3.5, I pointed out the issue with existing benchmark’s evaluation metric: it rejects perfectly…
-
What is Coming Next for Text-to-SQL
Text-to-SQL is a natural language processing (NLP) task that involves converting natural language questions into SQL queries that can be…
Subscribe with RSS.