Week of December 9th
Stefan Krawczyk
Agentforce. | Original Co-creator of Apache Hamilton & Apache Burr | Pipelines & Agents: Data, Data Science, Machine Learning, & LLMs
TL;DR:
#Hamilton release highlights: Better TypedDict support and modular subdag example
Office Hours & Meet ups for Hamilton & Burr.
Free 30 minute lesson using pytest for LLM work!
Blog -- Burr + BentoML: LLM agent deployment made easy
In the wild: Burr mentioned LLM Orchestration Ecosystem
Hamilton Release Highlights:
Hamilton == 1.85.0
The extract_fields decorator can now handle expanding TypedDict annotations! This is a quality of life improvement because it can reduce code you have to maintain and update. For example to extract and below from the output of `some_function, we can do the following:
Of course if one doesn't want to expose all the fields in the TypedDict one can provide a subset manually:
Modular Subdag Example
We were getting more questions around parameterization and reuse of Hamilton code. So we've added a modular machine learning type example that shows:
How to reuse common data transformation logic.
How to reuse common inference logic.
How to build a single Hamilton dataflow that has a "training" component that feeds into a "prediction" component, where inference is performed on the training set, as well as an unseen hold out set.
Here is what the code in the example creates:
If you're interested in see this, we invite you to check out the code here.
sf-hamilton-ui==0.0.16
We shipped a small fix to better display run time inputs that are captured by the graph.
Burr Release Highlights - stayed tuned for next week!
Office Hours & Meetup
Hamilton Meet up: Our next meet-up will be next week! We're excited that Jernej Frank will be presenting. Join/sign-up here.
Hamilton Office Hours: They happen most Tuesday 9:30am PT - 10:30am PT.
Join our slack for the link.
Burr Office Hours: They happen most Wednesdays 9:30am PT - 10:30am PT.
Join our discord for the weekly link.
Free Lesson on using pytest for LLM related work
I’m hosting a #free event with Hugo Bowne-Anderson on Maven about "Mastering LLM Application Testing". This 30-min session is for #SoftwareEngineers, #DataScientists & #MachineLearningEngineers who want to know:
how to use #pytest to test & evaluate their LLM based applications.
how to think systematically about improving the stability of their #LLM outputs.
Join LIVE on Monday December 16th @ 4pm PT!
RSVP here 👉: https://maven.com/p/a6f9bf/mastering-llm-application-testing
Burr + BentoML: LLM agent deployment made easy
Our new blog by Thierry Jean this week shows you how to combine Burr with BentoML! BentoML is a an open source serving framework. Meaning it's really easy to create a web-service endpoint that exercises your Burr code. Click below for the post, or find the source code here.
In the Wild:
Burr was spotted on a LLM Orchestration Ecosystem post! Excited to get recognized here!