top of page

TEAM

Quome Prompt->app

Jim Schwoebel, Gautam Prakriya, Lindsay Warrenburg, Marcos Ortiz, Monalisa Dutta

clear.png

About the Project:
We're working closely with Quome Inc. to develop a multi-agent system that creates and deploys prompt-to-app agents. This means you take take a prompt, build about 300 sub-prompts using chain-of-thought, and get back a deployable application on Quome Cloud as a docker container. This is a complicated multi-agent system composed of many sub-agents including prompt->product requirements docs agents, prompt->openapispec, prompt->frontend, prompt->backend etc. and many other types of agents, and there are literally thousands of experiments to do to make this a reality.

Our team has identified three key threads of work (which you can help join as a manager or data scientst):
1. Prompt Engineering - design and develop effective prompts to guide the behavior of our agent systems.
2. Post-hoc Analysis of App Artifacts - Examine output artifacts produced by our deployed agent systems to identify areas for improvement.
3. DevOps/Productionization of a Prompt->App Solution - Integrate our custom prompt agents into a production-ready workflow, ensuring reliable deployment and seamless integration with other systems.

Some ideas for work scope:
- Junior data scientist = join one chunk / thread of work with guidance of senior
- Senior data scientist = lead a key feature within one thread of work
- Staff data scientist = lead a thread of work completely

Work is open to part-time committed roles alongside a job search to build a resume and/or gain experience outside day job. All collaborators will need to sign NDAs to access proprietary code base and contribute.

Given we have a large scope of work to make this a reality, we are very excited to engage an interdisciplinary group of ERDŐS INSTITUTE stakeholders.

Screen Shot 2022-06-03 at 11.31.35 AM.png
github URL
bottom of page