Job Description

Location: Remote
Compensation: Phantom equity (early team)

Type: Full-Time or Contractor

🔥 About the Opportunity
Inkognido is building emotionally intelligent AI avatars and NPCs that respond dynamically across games, remember interactions, and evolve with players.
This means:
Personality systems
Emotional engines
Memory layers
Emergent behavior across worlds
And with this level of depth, safety, alignment, and predictable behavior become mission-critical.
The AI Safety Engineer ensures that all models behave reliably, respectfully, consistently, and in accordance with the rules of the Inkognido universe.
This isn’t a “content filter” role.
This is behavioral engineering + alignment strategy + guardrail architecture inside a multi-world gaming ecosystem.
If you're passionate about ensuring AI outputs remain safe, grounded, consistent, and emotionally coherent — you will thrive here.

🚨 Your Mission
You will be responsible for:
Preventing undesirable AI behavior
Building output guardrails
Maintaining tone & emotional alignment
Ensuring stability across thousands of NPC instances
Monitoring emergent personality patterns
Designing fallback + override logic
You make sure our AI remains safe, aligned, and predictable, even as it becomes more advanced.

🛠 Responsibilities
Design and implement safety guardrails for NPC + avatar AI behavior
Test for harmful, unsafe, or immersion-breaking outputs
Build filters, constraints, and behavior overrides
Work closely with prompt engineers to refine tone control
Develop monitoring pipelines for emergent behaviors and anomalies
Establish safety review protocols for new models or updates
Collaborate with backend engineers on routing logic for safe outputs
Document risks, mitigations, and safety guidelines

🧠 Requirements

Must Have:
Experience with LLM behavior testing or AI safety frameworks
Strong analytical ability to detect risky or inappropriate outputs
Understanding of prompt engineering and model control
Comfort designing safe failover logic
Clear communication and documentation skills
Ability to systematically test edge cases

Nice to Have:
Background in psychology, behavioral science, or linguistics
Experience with RLHF, model alignment, or preference tuning
Familiarity with content classification or moderation systems
Understanding of emotional modeling and character tone
Experience with RAG, embeddings, or memory architectures
This role is perfect for someone who understands both AI behavior and human behavior.

💡 Why This Role Matters
Because we’re not building tools — we’re building characters.
Characters with:
Emotional states
Memory recall
Personalities
Reactive dialogue
Player influence
You are responsible for ensuring:
No harmful outputs
No breaks in emotional consistency
No immersion-breaking behavior
No unsafe interactions
No “rogue personality drift”
Your work protects the entire platform.
This is a central, high-trust role in the AI division.

🏆 What You Get
Phantom equity (early-stage pool)
Ownership of the AI safety pipeline
Freedom to design alignment systems from scratch
Direct collaboration with founders + lead engineers
The chance to define safety standards that scale to millions of players
Deep involvement in a space where safety is often ignored — but critical

✉️ How This Role Works With Others
You will collaborate closely with:
AI Lead Engineer – Safety constraints + behavior logic
LLM Engineers – Output alignment + routing
Prompt/Dialogue Engineers – Tone reinforcing + emotional guardrails
Memory/Emotional Engine Engineer – How memory influences behavior
CTO – System-level guardrail architecture
Realms Team – Safety standards for NPCs inside live gameplay
Your influence touches every interactive AI system.

📚 Study Links (Internal Use)
(Not visible to applicants unless desired)
AI Alignment Fundamentals
Behavior Constraint & Routing Techniques
Emotional Modeling for NPCs
Safety Testing for LLMs
Guardrail Architectures & Content Filtering Models

📩 How to Apply
Submit your application:

Or email:

[email protected]

[email protected]

[email protected]

[email protected]
Include:
Resume / LinkedIn
Any AI alignment or safety work you’ve done
Prompt examples or behavioral tests (if available)
A brief note describing your approach to safe AI
We evaluate every applicant manually.

Job Tags

Similar Jobs