I’ve been working on optimizing AI agent working memory this past week.
Did a lot of research, and our team decided to hold a study group where I’ll present what I learned about AI agent memory.
Here's that presentation.
hope it’s useful for you:
https://t.co/RBcSVRUfNH
Almost anything I read on X is shallow. Another amazing AI tool, another new agent, and … Yes, they are really amazing
But why does nobody talk about the business itself? 90% of these projects fail. Why? I think it is because nobody cares about the core and the actual product.
Step 3: Retrieval Skills
Learn to bring external knowledge into your models using embeddings and semantic search — the foundation of RAG (Retrieval-Augmented Generation).
Ever felt like AI is evolving faster than your ability to keep up? 😬
You see new AI products every day and think “I’ll just use an API from OpenAI.”
Yeah… no. It’s way deeper than that. Here’s what you actually need to learn 👇
Step 2: Prompt Engineering
This is where the magic happens.
Writing prompts and crafting context that trigger *the best possible responses* is both art and science. Master this, and you’ll stand out
Amazon didn’t boost sales with a new feature.
They just changed one word.
“Sign up” became “Continue.”
People stopped feeling like they were starting over and started finishing what they came to do.
Debugging in the dark? Here’s your flashlight.
My latest guide breaks down:
Manual vs. automatic instrumentation
How Sentry uses monkey patching
Tips to catch Node.js issues early
Fewer surprises in prod 👇
https://t.co/fSSkKP3Wjc
@denicmarko I provide consulting & mentorship for anyone looking to level up their career.
✅ Resume reviews & professional resume writing
✅ Interview prep & guidance
✅ Mentorship for software engineers aiming to become AI engineers
DM if you’re interested