Femi Ad ๐๐ฅ
36K posts
Femi Ad ๐๐ฅ
@Hallengray
Black. Proud. Gifted. Entrepreneur | AI Software Engineer | AI Automation | Restaurateur ๐ฅ @KaaboLounge Building ideas, telling stories, & automating โค๏ธ MUFC
London, England Joined Ekim 2009
768 Following1.4K Followers
He is a mumu man.
UtdDistrict@UtdDistrict
๐จ | Andrรฉ Onana's current mindset is that he will be RETURNING to #mufc, hoping to play during pre-season, and convince the new manager he can be part of his plans. Onana still believes he has a future and is not prepared to give that up without a fight. [@sistoney67]
English
RAG-Forge v0.2.1 is out:
โข Safety refusals scored fairly
โข No more silent zero-scores on broken measurements
โข Brand-new audit report (NIST AI RMF + ISO 42001 + EU AI Act ready)
MIT. Framework-agnostic. Drop-in CI gate.
npm i -g @rag-forge/cli@latest
Femi Ad ๐๐ฅ@Hallengray
v0.1.3 just shipped: 6 packages on npm + PyPI MIT licensed Audit-as-CI-gate workflow MCP server included โญ github.com/hallengray/ragโฆ ๐ฆ npm i -g @rag-forge/cli ๐ rag-forge-docs.vercel.app DM if you want a free RMM audit on your stack.
English
Ran a production RAG audit for a client yesterday.
Their system was passing every test they had. It was also getting punished for doing the right thing โ every time it safely refused an unsafe question, my evaluator scored it a zero.
So I fixed my evaluator.
Femi Ad ๐๐ฅ@Hallengray
v0.1.3 just shipped: 6 packages on npm + PyPI MIT licensed Audit-as-CI-gate workflow MCP server included โญ github.com/hallengray/ragโฆ ๐ฆ npm i -g @rag-forge/cli ๐ rag-forge-docs.vercel.app DM if you want a free RMM audit on your stack.
English
If you are interested in RAG. #llm
dev.to/hallengray/theโฆ
English
Typescript and lint in dev env = PASS
CI in Github = FAIL
My head!
GIF
English
Just hit 55k! You can try the most recent and more powerful version built on @TwelveData 's API.
n8n.io/workflows/1398โฆ
Femi Ad ๐๐ฅ@Hallengray
My AI stock analyst n8n workflow just hit 48K+ views. Built an intelligent Telegram bot that analyzes stocks like a 50-year NYSE veteran using Claude 3.5 Sonnet and GPT-4O Vision. No backend. No complex infrastructure. Just n8n automation. #n8n #automation
English
How I start my day - Open my eyes ใ Sit on my desk ใ CLAUDE on my CLI
English
v0.1.3 just shipped:
6 packages on npm + PyPI
MIT licensed
Audit-as-CI-gate workflow
MCP server included
โญ github.com/hallengray/ragโฆ
๐ฆ npm i -g @rag-forge/cli
๐ rag-forge-docs.vercel.app
DM if you want a free RMM audit on your stack.
English
Most production pipelines I've audited sit at RMM-0 or RMM-1.
Until you measure where you actually are, you don't know what to fix next.
The RMM gives you a roadmap, not just a score.
English
Most RAG projects ship without ever knowing if they actually work.
I built RAG-Forge to fix that.
A framework-agnostic CLI for production RAG with evaluation baked in, not bolted on after deployment.
Here's what it does โ
English
Most RAG projects ship without ever knowing if they actually work.
I built RAG-Forge to fix that.
Framework-agnostic CLI: scaffolds pipelines, runs eval as a CI gate, scores any system 0โ5 against the RAG Maturity Model.
v0.1.3 live: github.com/hallengray/ragโฆ
English