About the job
About Pax Historia
Pax Historia is pioneering a new genre of gameplay by leveraging cutting-edge generative AI technologies. Our innovative platform combines the strategic depth of grand strategy games with the limitless creativity of a sandbox environment, driven by a vibrant community that actively creates and modifies scenarios.
Our user base generates hundreds of scenarios daily and engages in millions of game rounds each week, exhibiting rapid growth. We are proud to be supported by esteemed investors such as Y Combinator, Pace Capital, and Z Fellows. Your contributions will have an immediate impact on a product enjoyed by hundreds of thousands of players.
Position Overview
We are seeking a founding-level ML Systems Engineer to join our team in-person full-time in San Francisco (Dogpatch). You will have the opportunity to work closely with our cofounders to shape the future of our technology.
Current Challenges:
- Closed-source models yield satisfactory game performance but come with high costs.
- Open-source models are more budget-friendly yet often underperform in our environment.
- Prompts and harnesses exhibit minimal variance across models.
- We have a functional internal evaluation system with significant opportunities for enhancement.
Your Responsibilities:
- Establish and manage the necessary infrastructure for customizing harnesses and prompts tailored to individual AI models to optimize their performance.
- Develop domain-specific models aimed at narrowing or completely bridging the gap in performance between open and closed models.
- Optimize caching strategies to decrease expenses associated with closed-source models.
- Enhance the performance of closed-source models by training specialized endpoints.
- Assess and advance the effectiveness of embedding and reranking mechanisms in our applications.
- Facilitate the creation of new user experiences based on emerging world models.
In Summary: Your work will directly enhance the affordability and enjoyment of the game.

