Lead AI / LLM Systems Engineer – Egofold

Snail Games USA•Culver City, CA

4h•$150,000 - $200,000•Hybrid

About The Position

Egofold is an AI initiative within Snail Games focused on building a modular AI “brain” ecosystem for NPC intelligence, real-time perception systems, and simulation tooling across multiple game projects. We are seeking a senior AI engineer to lead the development of Egofold’s foundational AI systems. This role is responsible for building and structuring a reusable, trainable LLM-based “brain” that can operate across multiple environments and contexts. The focus is on core system design and execution: how the AI reasons, learns, adapts, and integrates training workflows over time. This is a hands-on role for someone comfortable making technical decisions under ambiguity and operating without fully defined requirements.

Requirements

Significant professional experience building AI or ML systems beyond simple model or API integration.
Demonstrated experience working with large language models in a production or applied research context.
Hands-on experience with agent training methodologies, including reinforcement learning or simulation-based learning systems.
Strong understanding of training workflows, evaluation strategies, and iterative improvement cycles.
Strong proficiency in Python and experience integrating AI systems into production environments.
Ability to reason about complex, stateful systems and learning behavior over time.
Comfort operating in early-stage, ambiguous environments and taking ownership of foundational systems.

Nice To Haves

Experience designing context-aware or agent-based AI systems.
Background in behavioral AI, simulation, or decision-making systems.
Familiarity with reinforcement learning, fine-tuning strategies, or hybrid AI architectures.
Experience integrating AI systems into real-time or interactive environments.
Familiarity with C++, C#, or Unreal Engine integration workflows.

Responsibilities

Design and implement foundational AI systems using large language models as a core reasoning component.
Architect simulation-based learning systems and training workflows, including fine-tuning, reinforcement learning, evaluation, and feedback loops.
Build validation, safety, and constraint layers around generative outputs to ensure predictable and controllable behavior.
Define evaluation frameworks and benchmarking strategies to measure agent performance, stability, and learning progression over time.
Structure how context, memory, and world state are represented and consumed within the AI architecture.
Determine how learned behavior, structured logic, and rule-based systems interact within a unified hybrid system.
Collaborate with engine engineers to integrate AI systems into real-time interactive environments.
Make system-level technical decisions that prioritize long-term reuse, scalability, and cross-domain adaptability.