Notes from building AI agents that get measured.
Mostly evaluation, agent and retrieval architecture, and the failure modes that only show up once real users are asking real questions.
A Playbook on Considerations when building AI Applications in Production
Why most AI demos die in production — and the framework that gets them to survive
Write-Audit-Publish (WAP) pattern for robust data pipelines
This is a guide on how to implement a WAP pattern in snowflake and dbt
How we cut our Snowflake bill in half overnight
Controlling Snowflake Costs: Recommendations and Common Pitfalls to Avoid
Determining the Necessity of MLOps in the SaaS Tech Industry
A Comprehensive Framework
LLM Fine-Tuning: A Hands-On PEFT Lora Example
Utilising PEFT's (Parameter-Efficient Fine-Tuning) Lora (Low Rank) method to fine-tune a LLM.
Agentic Workflows: Why AI Agents are on the rise
Knowledge Assistants are supercharging RAG and allow the use of tools.
Understanding Temperature, Top-P, and Top-K in AI Models
A hands-on example based on Meta Llama
Harnessing NLP for Sentiment Analysis in Customer Support
A HuggingFace example leveraging the Twitter-roBERTa-base model
Mistral: Creating an Intelligent Chatbot with LangChain
A Comprehensive Guide to Building and Deploying a Local AI Chatbot Using Mistral and LangChain
Unlocking the Power of LangChain and AWS Bedrock for Retrieval-Augmented Generation (RAG)
Let's add a file as context for the LLM.
Using AWS Bedrock to build an AI agent in Python (Boto3)
An example on how quickly you can get started with AWS Bedrock to build a AI application.
Working on something in this territory?
I take on a small number of builds and audits at a time. A short description of the failure you're seeing is enough to start.