Member of Technical Staff, AI Researcher (New Stealth Startup)
Redcellpartners · Remote (USA)
About this role
## About Us Red Cell Partners is an incubation firm building and investing in rapidly scalable technology-led companies that bring revolutionary advancements to market in three practice areas: **healthcare, cyber, and national security**. We’re united by a shared sense of duty and a deep belief in the power of innovation. ## About the Stealth Startup We’re launching a new startup redefining how mission-critical organizations adopt AI. Our platform helps organizations across **healthcare, government, and other regulated industries** transform institutional knowledge into intelligence assets—improving the **accuracy, reliability, and impact** of AI in real-world workflows. We’re a fast-moving team of builders, researchers, and operators helping customers **create, measure, and continuously improve** AI capabilities they can own and control. ## About the Role We’re seeking a **Member of Technical Staff, AI Researcher** to help turn complex, real-world workflows into the **evaluation frameworks, reinforcement learning environments, and feedback systems** needed to advance frontier AI. This is a blended research and engineering role. You’ll investigate how **data, evaluations, and feedback systems** influence model behavior, while building the infrastructure and tools required to test and apply those ideas. ## Responsibilities - Design and build **evaluation frameworks, benchmarks, and methodologies** for frontier LLMs, AI agents, multimodal models, and reinforcement learning environments - Develop **reward models, programmatic verifiers, graders, and feedback systems** to make model behavior measurable and improvable - Design and run controlled **supervised fine-tuning and reinforcement learning experiments** to measure effects of datasets on capability, generalization, and alignment - Research and define what makes evaluations **representative of real-world work** and operational environments - Build systems for producing **high-quality human data**, including expert task design, annotation methodologies, data-quality signals, and data-attribution techniques - Translate research questions into **lightweight, rigorous experiments** that generate actionable findings - Prototype quickly, evaluate results, diagnose failure modes, and convert findings into improved **benchmarks, datasets, or technical systems** - Move promising methods from experimentation into **scalable, production-quality implementations**
Listing freshness
CronJobs last confirmed this listing 2h ago. If its source stops confirming the opening for seven days, this page is removed from active inventory.