AI Consulting
Is your business using AI but are feeling stuck? My clients often feel overwhelmed navigating the sea of options and trade-offs when rolling out AI and often just want to know if they're doing it right.
As a hands-on AI architect, I show where you're doing great, as well as provide actionable feedback on where you can improve as well as give ideas on where you can go next. I have 17+ years of experience in software engineering, architecture and management, as well as 6+ years producing and operating AI/ML applications.
- AI architecture — RAG, knowledge graphs, vector databases, etc.
- AI/ML Operations — Deploying, testing, and monitoring AI or ML apps.
- Rollout — ChatGPT or Microsoft Copilot
- Engineering — Using AI code generation tools effectively
- Education — Programs that enable employees to get the most from AI
Interested? Contact me for a first consultation. I prefer longer-term engagements that go beyond initial advice.
Contact MeRelevant Articles
Who Wins With Cursor & Copilot?
Who wins now that Cursor is out? Good programmers? Bad programmers? Here, I make the case that it is more about your tempermant and personality traits. If you're good in the places where the AI is weak, you'll knock it out of the park.
Does Prompt Caching Make RAG Obsolete?
Anthropic announced prompt caching today. How is it helpful? Does it replace RAG? Let's discuss.
Vector Stores Are Dumb
Vector stores are used to implement the RAG pattern, but they're of limited utility. In this post I explain why I think vector-based RAG is limited and why using knowledge graphs to implement RAG is better.
Accountants Should Do Hackathons!
Everyone should do hackathons! Magic happens when you empower the people with the problems to solve their own problems.
Why The Llama 3.1 Announcement Is Huge
Today Meta announced Llama 3.1 405B as an open source AI model. Here's why this is a game-chager for the AI landscape.
RAG Trick: Embeddings are Spheres
This post offers some helpful simplifications you can make when working with RAG or embeddings that help build a working mental model around them. Embeddings typically form a (hyper)sphere, because they're normalized.
LLMs are Interpretable
Shockingly, LLMs are the most interpretable form of machine learning that I've seen so far, in that it's very compatible with an end user's needs for trust and explanation of behavior.