About the Role At Something Else, we're building next-generation AI agents that reason, adapt, and learn autonomously in real-time environments. Our mission is to harness the transformative potential of AI to further human knowledge and elevate global well-being. We believe that agents capable of understanding, exploring, and interacting meaningfully with the world can help solve some of humanity's most complex challenges.
As an AI Agent Engineer, you'll be responsible for designing and implementing scalable, cognitive-agent frameworks that embed planning, memory, and real-world interaction using state-of-the-art language models and symbolic systems.
Focus- Develop modular agentic systems with plug-and-play cognitive functions (e.g., memory, planning, metacognition).
- Integrate language models into active environments with real-time perception and feedback loops.
- Create benchmarks and simulations that test multi-agent interactions in open-world settings.
- Apply reinforcement learning, active inference, or planning algorithms to drive emergent behavior.
Ideal Experiences- Strong foundation in cognitive architectures (e.g., Soar, ACT-R, ReAct, Tree of Thought).
- Experience building agent-based simulations or environments (e.g., OpenAI Gym, Habitat).
- Familiarity with real-time decision-making under uncertainty (e.g., POMDPs, Bayesian approaches).
- Deep proficiency in Python, PyTorch/Jax, and vector search.