Vibeleaderboard
← All Intel
Intel / video

Stanford CS229 I Machine Learning I Building Agents That Do the Work of Human Software Engineers

Source
youtube.com
Author
Stanford Online
Date

Context

In a Stanford CS229 lecture, speakers building agents that do software engineers' work ran a live demo of an investigating a production problem across many tools. As narrated, when told a user interface was slow, the agent mapped the question to the front-end service, searched monitoring, infrastructure and code tools, and returned charts with cited spikes. Asked for a root cause, it reported that a feature flag had been switched on by a scheduled job as a deliberate failure test.

In a second run, given only that an was slow, the agent wrote its own follow-up questions for other agents and tools. The speakers' motivation is that engineers spend most of their time keeping systems reliable, with humans acting as glue between tools. Because this was a presenter-run demo of the speakers' own system, it shows the design and one successful investigation, not how often such agents find the right cause.

Terms in this piece · Glossary
  • Agent — A system in which a model repeatedly chooses operations, a harness carries them out, and the results return to the model to inform its next choice.
  • Latency — The time between asking for a result and having it, including the model's generation time, tool calls, retries and any queueing.
  • API — Application programming interface: the agreed set of operations, inputs and outputs through which one piece of software uses another.
Read the source www.youtube.com
More from Stanford Online
Recommended reads
Comments

Checking sign-in…

Loading comments…