luisjunco

🔴 Red Team Challenge

Red teaming is the practice of deliberately trying to break, misuse, or bypass a system so its weaknesses can be found and fixed before they cause real problems.

In AI engineering, this means going beyond asking “Does the model work?” and asking:

• “How could someone misuse this?”
• “What assumptions is the developer making?”
• “What happens when the user doesn't follow the intended path?”
• “Can I make the model ignore its instructions?”

luisjunco

NLP Evaluation Metrics

1. Intro to BLEU (machine translation)

Watch this video:

• "What is the BLEU metric?" (5 minutes): https://www.youtube.com/watch?v=M05L1DhFqcw

luisjunco

Open Models vs. Closed Models

Open-weight models (e.g. Kimi3, Llama, Mistral, Phi)

Advantages

• Can fine-tune on your own data
• More privacy and control (can run on your own infrastructure)
• No per-token API costs (you pay for compute instead)
• Can run offline or in air-gapped environments
• Full control over model version and deployment

luisjunco

LangChain Versions: Evolution, Terminology, and Key Differences

LangChain has evolved significantly since its initial release. Rather than thinking in terms of dozens of individual releases, it's more useful to understand its development as three major eras.

Era	Common name	Approx. timeframe	Main characteristics
Pre-0.1	Early LangChain	2023	Experimental framework centred around chains
0.1–0.3	LangChain v0 / LCEL LangChain	2024–2025	Introduction of LCEL and the Runnable architecture
1.x	LangChain v1	2025–present	Agent-first framework built on top of LangGraph

luisjunco

Practice: LangChain Agents

Setup

Install dependencies:

#

luisjunco

Core AI Engineering Techniques & Patterns

Model Development & Adaptation Techniques

luisjunco

Practice: From Prompts to Chains

Iteration 1 — Prompt template + simple chain

Create a chain that translates a word into a given language.

1. Build a PromptTemplate with two input variables, word and language.

luisjunco

Practice: Building Chains

You've learned how to combine a PromptTemplate and an LLM into a chain using the | operator, then run it with .invoke().

Iteration 1 — Environment Setup

• Create a notebook (e.g. locally on VS Code, or on Google Colab).

luisjunco

Machine Learning - Model Deployment Demo - STEPS WE'LL FOLLOW

Intro to Deployment

• Environments (dev / staging / production)
• Recap: what's an API
• Architecture

luisjunco

Machine Learning Project Structure: Recommended Practices

This repository structure is a recommended way to organize machine learning projects.

It helps keep notebooks focused on exploration, moves reusable code into Python modules, and makes projects easier to understand, maintain, and reproduce.

Feel free to adapt it to your project's needs —there is no single "correct" structure.