Built an agentic document review system for vendors contracts, amendments, invoices and other documents that describe the same business transaction.
Strong project added to resume๐ช๐ป
Built a little research agent today๐ค
It can crawl through a codebase, spot potential issues, and uses RAG to dig through Next + React knowledge when it gets stuck.
Basically gave my agent a codebase and said: โgo solve these issuesโ๐
Very simple, though was a fun challenge.
Why though right?
Because LLM are frozen in time. Update to libraries or code is not on their training data.
Neither they have private knowledge.
So we take the data -> store in vector DB
Retrieve when needed. Simple๐ค
RAG - Retrieval-Augmented Generation
(Its in the name literally lol๐)
Instead of making the LLM memorize everything, you give it a search button and let it grab the relevant context when it needs it.
LLM: โI donโt know.โ
RAG: โhold on, lemme look that up.โ ๐
Introducing Memorable (YC S27): PROCEDURAL MEMORY FOR AI AGENTS
AI agents today are born, work, and die inside a single context window. They solve a hard problem once, then start from zero when it returns.
Memorable turns successful runs into a graph of reusable procedures.
So every task makes the next one:
- Faster.
- Cheaper.
- More deterministic.
Weโre excited to share that Memorable is joining Y Combinatorโs S27 batch.
Try it: https://t.co/DUAVdejTzZ