RegIntel — What Does an AI/LLM Engineer Actually Do? RAG, Knowledge Engineering, Context - Part-8

Описание: What are AI/LLM Engineers actually doing when they join a project? What tech do they use? What are the key activities?

In this video (Part 8 of 8), I explain AI/LLM Engineering through the RegIntel project — how we apply intelligence on top of regulatory knowledge, and what the engineer's day-to-day work actually looks like.

What I cover:
→ AI/LLM Engineers do the tech work — but the domain comes first, technology comes second
→ Like a carpenter has tools but you tell him the size of the sofa — same with AI engineers and the business problem
→ Knowledge Engineering: structuring PDFs, books, and regulatory documents into a queryable system
→ Context Engineering: when a user sends a query, we match it with the right context from the right document
→ The LLM receives the query + context and generates a cited answer — that's RAG (Retrieval Augmented Generation)
→ Graphs come in when information gets too large — hundreds of documents need structure, not brute force
→ Real example: we built a working prototype in 5-6 hours — Node.js APIs, Cheerio scraping, PDF parsing, OpenAI embeddings, WebSockets, custom recursive text splitter
→ Why? Because we understood the problem. We'd been studying it for months. Execution was fast because clarity was already there.
→ My suggestion: don't pick a technology first and search for a problem — pick the problem first, then find the technology

First pick up the problem. Then go to the solution.

🔗 RegIntel Home: https://getpostlabs.io/insights/regte...
🔗 Part 3 — How RegIntel Works (KE Layer): https://getpostlabs.io/insights/regte...
🔗 Part 4 — Tech Architecture (AI/LLM layer): https://getpostlabs.io/insights/regte...

---
🔗 GetPostLabs: https://getpostlabs.io
🔗 LinkedIn:   / sumitarora