SlipstreamJobs tracks this role from the company's public career site. Apply directly on the employer's site.
Ineffable Intelligence is building a superlearner AI system that discovers knowledge through reinforcement learning, with the goal of achieving superintelligence. The company's mission is to create systems that can rediscover and transcend major human intellectual achievements like language, mathematics, and science.
As a Software Engineer for Simulated Universes, you will design, build, and maintain the virtual environments in which these superlearners train and develop. This is a foundational engineering role critical to the company's core mission. Your responsibilities include:
- Architect and implement the engineering foundations of new simulation environments, ensuring they are accurate, reliable, and intricate
- Work closely with researchers and designers who define the specifications for these worlds, translating their blueprints into functional systems
- Enable rapid research iteration by building codebases that support fast experimentation and efficient execution
- Scale environments to unprecedented levels of parallelism and speed using state-of-the-art hardware
- Maintain and evolve systems as they grow in complexity and scale
You will be working at the intersection of cutting-edge AI research and systems engineering, partnering closely with the research team to turn ambitious visions into reality.
REQUIREMENTS:
- Industry experience with large-scale simulations (scientific domains or games industry)
- Demonstrated expertise building distributed systems at scale (e.g., MMO architectures, world parallelism, large-scale physics engines)
- Ability to design and maintain codebases that support both fast research iteration and efficient execution
- Experience working with physics engines and GPU-accelerated simulation frameworks (NVIDIA Warp or similar)
- Proficiency in lower-level languages: Rust, C, or C++
- Experience implementing distributed server-client architectures
- Capability to write efficient GPU code
- Comfort working closely with researchers and designers, translating technical requirements into systems
