Build
Built an agent that triages my GitHub issues into a Kanban every morning. 40 lines + a cron.
Lena Hoffman
@datadruid
Data viz tinkerer. Into #dataviz and #machinelearning, building with @shipfast_sara. Notes at https://example.com
Ships 3 creations recently · mostly build + question posts · into Agents, Tutorials, LLMs
researcher
Joined June 2026
- Posts
- 8
- Remixes
- 1
Creations
3 creationsShip graph
· last 12 months9 in 6 days8 ships and 1 accepted answer across 6 active days in the last year. Busiest: 2 ships on Jun 30, 2026; 2 ships on Jun 11, 2026; 2 ships on Jun 7, 2026.
More about @datadruidShowHide
Verified proficiency
- Resolved a question
Achievements
- First Ship
- Forked
- Accepted Answer
Featured Build Clips
Fast walkthroughs, demos, and shipping proof from @datadruid.
DuckDB + an LLM = ask your CSV questions in English and get a chart back. Here's the 30-line bridge.
Open clip post →Posts
Question
Best approach to dedupe near-identical embeddings before upserting to a vector DB? Cosine threshold feels brittle across domains.
Replies
Thought
Re-embed. Embedding spaces from different models aren't aligned — a linear projection between them loses exactly the nuance you switched models to get. Batch the backfill: export ids + text, run it through the new model at off-peak rate limits, and write to a second column/index…
Thought
What's your eval setup for this? I want to steal the deadlock-detection bit for my pipeline.