Junnn

3 ms·

1. 1.▲
   Show HN: DataFlow,Turn raw data into high-quality LLM training datasets(github.com)

   2 points by Junnn 7mo ago|0 comments
2. 2.▲
   by Junnn 7mo ago

   I’ve only used it on my computer.
3. 3.▲
   Show HN: Generate, Clean, and Prepare LLM Training Data, All-in-One(github.com)

   2 points by Junnn 7mo ago|0 comments
4. 4.▲
   Show HN: Synthesize complex agent training data with just a few lines of code(github.com)

   1 points by Junnn 7mo ago|0 comments
5. 5.▲
   by Junnn 9mo ago

   From an engineering perspective, what I find compelling here is not “no embeddings”, but the decision to treat memory as a first-class, inspectable system rather than a retrieval trick. Most agent memory stacks today collapse everything int
6. 6.▲
   by Junnn 9mo ago

   I’m working on a sales assistant agent with long-term memory. What database does memU support by default? I’m using pg.
7. 7.▲
   by Junnn 10mo ago

   This is really helpful — thanks for sharing. If you don’t mind me asking, how long did it take you to build this? （I’m thinking about creating a small aggregation site to track interesting Agent projects and would love to get a sense of the
8. 8.▲
   by Junnn 11mo ago

   I’d like to ask why there is a ‘reply’ section — it seems to repeat the previous sentence.
9. 9.▲
   by Junnn 1y ago

   wow