-
Building Safer AI Agents: 8 Essential Concepts for Tool Execution Safety and Reliability
AI agents are becoming increasingly capable of interacting with external tools, APIs, databases, and services. But giving an LLM access to tools introduces an important engineering challenge: How do we…
-
Building a Tokenizer From Scratch: Byte-Pair Encoding, Verified on Real Data (Ep:00.02)
Whole-word tokenization hits a wall the first time it meets a word it’s never seen. This episode builds Byte-Pair Encoding from raw Python, trains it on the real TinyStories corpus,…
-
Colab Setup + TinyStories — Picking the Right Dataset to Train an LLM From Scratch (Ep:00.01)
Set up a free Google Colab environment for training an LLM from scratch, and learn why dataset choice matters more than model size when you’re training small. We load TinyStories…
-
From Scratch to Agents — Build a Real LLM, Fine-Tune It, Add RAG, Make It an Agent (Ep:00.00)
A hands-on, code-only course: build a real GPT-style LLM from scratch on Google Colab’s free tier, pretrain it on real text, fine-tune it with LoRA, add RAG, and turn it…
-
Multi-Agent Systems Explained: A Sub-Agent Is Just a Tool That Runs Its Own Loop (Ep:06.08)
An orchestrator agent invoking a specialist sub-agent is structurally identical to a single agent calling a tool — the only difference is that “executing” the action means running an entire…
-
Why “More Training” Didn’t Fix the Undertrained Tool — a Deeper Diagnosis (Ep:06.07)
Oversampling a tool’s training data 4x more didn’t fix it — because the real problem wasn’t repetition count, it was the size of the tool’s input domain relative to what…
-
Why Self-Correction (Reflexion) Doesn’t Fix an Undertrained Model (Ep:06.06)
Reflexion lets an agent verbally reflect on its own mistakes and retry, entirely in-context, with no weight updates. Module 6 of From Zero to Agents tests it directly on a…
-
Fixing a Broken Evaluation Metric: Comparable Positions, Not Comparable Sequences (Ep:06.03)
The fix for a per-tool evaluation metric confounded by trace length isn’t more data — it’s evaluating only the one position that’s genuinely comparable across every tool. Module 6 of…
-
Catching an Undertrained Tool Before It Fails: Two Signals, and a Real Evaluation Trap (Ep:06.02)
Per-tool held-out loss and label-free generation confidence both aim to catch an undertrained agent tool automatically — but one of them walks straight into the exact evaluation-scope trap Module 5…
-
Multi-Tool Agents and Chained Tool Calls: Extending the ReAct Loop (Ep:06.01)
Choosing between tools is just next-token prediction over a bigger action vocabulary, and chaining tool calls is the same loop run more than once. Module 6 of From Zero to…