We’re working with an early-stage AI startup tackling one of the biggest hidden problems in gaming.
Every year, game studios spend countless hours on manual QA, repetitive testing, bug reproduction, and validation. This team believes developers should be spending their time building worlds, not chasing bugs.
Their mission is to build autonomous AI agents capable of understanding and interacting with games much like a human player would.
The company is still small, highly technical, and moving fast. They’ve already developed proprietary AI models that outperform some of the biggest names in the industry on key benchmarks, and they’re now looking for a Founding AI Engineer to help push the next generation of game-understanding systems into production.
The Role
This is not a research lab role. It is not an API-wrapper role either.
You’ll sit at the intersection of cutting-edge AI research and real-world product development, owning core perception systems that help autonomous agents understand what is happening inside games.
You’ll work directly on Vision-Language Models, video understanding, temporal reasoning, and production deployment.
Your work will become the foundation upon which the company’s autonomous testing agents operate.
What You’ll Be Working On
* Fine-tuning and optimizing Vision-Language Models (VLMs)
* Building systems that understand gameplay from images and video
* Designing approaches that capture temporal context across multiple frames
* Training and evaluating models on large-scale gameplay datasets
* Improving inference speed, latency, and deployment efficiency
* Taking models from experimentation all the way into production
* Working closely with founders to shape the long-term AI roadmap
We’re Looking For Someone Who
* Has hands-on experience fine-tuning Vision-Language Models
* Has worked extensively with video data and temporal modeling challenges
* Enjoys applied AI and shipping production systems
* Can move comfortably between research and engineering
* Understands the trade-offs between accuracy, latency, scalability, and cost
* Has experience deploying ML systems into real-world environments
You may currently be an:
* AI Engineer
* Applied AI Researcher
* Machine Learning Engineer
* Computer Vision Engineer
* Multimodal AI Engineer
What matters most is that you’ve built things, not just studied them.
Nice To Have
* Experience with models such as LLaVA, CLIP, Flamingo, Gemini, Qwen-VL or similar architectures
* CUDA optimization experience
* PEFT, LoRA, quantization, or model compression techniques
* Real-time inference optimization
* MLOps and deployment experience
* Experience working with multimodal foundation models
You’ll Probably Enjoy This Role If
* You like solving difficult technical problems with real-world constraints
* You want ownership rather than a narrow specialization
* You enjoy working closely with founders
* You care about building products, not just publishing papers
* You want to help define the technical foundation of a company from an early stage
Not The Right Fit If
* You’re only interested in academic research
* You prefer highly structured corporate environments
* You avoid production responsibility
* Your experience with VLMs is mostly theoretical
This is an opportunity to join very early, work on genuinely challenging AI problems, and help build the future of autonomous game testing.