Harness Engineering, the first frontier!

​Harnesses are understudied and the LLM labs' best-kept secret.

​Those GPT wrappers hide a lot more complexity and control. We look at how this shapes new "backend engineering" with LLMs as the 4th primitive (compute, network, storage).

​We talk about the coding harness as the most general harness, where all innovation lands. But the case for narrow harness remains.

​Nirant will be talking about Claude Code and Hermes as the harness to leverage the model intelligence

​About the speaker:

​Nirant, founder of Scaled Focus, help AI teams ship agents that stay reliable in production — build, monitor, and eval.

​Nirant Kasliwal is a notable AI Engineer with over 7 years of experience in areas such as chatbots, language models, and vector databases.

​He has contributed to AI education through projects like "Awesome NLP," a resource for engineers learning NLP and publishing NLP with Python.

​To attend online:

• ​⁠Add to calendar: https://shorturl.at/5uV9G

• ​Gmeet link: meet.google.com/gzr-tfdk-kyr

​Looking forward to seeing you!