Computer Vision Data Annotation Specialist
Job Description
Role Title: Computer Vision Data Annotation Specialist
Role Type: Contractor
Location: Northern America, Latin America, Europe
micro1 is engaging Computer Vision Data Annotation Specialists to contribute to a project advancing hand tracking and pose estimation technologies. In this role, you'll apply your expertise to help train next-generation AI systems. Your work will shape how models learn, reason, and perform through high-quality, real-world input. No prior experience in AI is required — your domain knowledge is what matters.
Scope of Work
Review and annotate images and video frames, focusing on detailed hand pose landmarks including wrists, finger joints, and fingertips.
Validate the accurate identification of left and right hands in diverse scenarios.
Detect, flag, and document edge cases such as occlusion, motion blur, gloves, multiple hands, and low-light conditions.
Apply thorough annotation guidelines and quality rubrics to ensure data reliability and consistency.
Maintain high accuracy and productivity when managing large datasets for computer vision training.
Report ambiguous, challenging, or unclear cases to project leads for clarification and guidance.
Perform quality assurance reviews on previously annotated data as needed.
Preferred Qualifications
Proven experience in image or video annotation for machine learning or AI projects.
Proficiency with annotation platforms.
Strong attention to detail, spatial reasoning, and ability to follow intricate visual guidelines.
Basic familiarity with computer vision concepts such as landmarks, bounding boxes, confidence scores, and ground truth labeling.
Experience with hand or human pose estimation datasets, or frameworks.
Background working on gesture recognition, sign language, AR/VR hand-tracking, or related datasets.
Excellent written and verbal English communication skills for documenting findings and collaborating remotely.
About Micro1
Micro1 is building the essential infrastructure for the next generation of artificial intelligence, moving beyond the limitations of static datasets to create a dynamic ecosystem where models truly learn to think and act. At its core, the company understands that the path to frontier intelligence isn't paved with synthetic data alone; it requires the nuance, judgment, and contextual awareness that only expert human guidance can provide. By positioning itself at the intersection of human expertise and advanced reinforcement learning, Micro1 is tackling the hardest problem in AI today: teaching models not just to answer questions, but to reason through complexity and take meaningful actions in unpredictable, real-world environments.
This mission comes to life through Realm, Micro1's flagship training environment. Realm functions as a high-fidelity simulation ground where AI agents are immersed in scenarios that closely mirror the messiness of actual human workflows. Rather than relying on abstract benchmarks, Realm forces models to engage in agentic actions, multi-step decisions, digital navigation, coding, and problem-solving that require genuine comprehension. It is within these realistic sandboxes that world-class human data is generated, creating a continuous feedback loop where expert annotators observe, correct, and guide model behavior. This process does more than just fine-tune outputs; it fundamentally elevates a model's underlying reasoning architecture, ensuring that the intelligence developed in training holds up when deployed into the wild.
But training is only half the equation. Micro1 recognizes that the true test of an AI system lies in its production performance, which is where Cortex enters the picture. Cortex is a contextual evaluation platform designed to move beyond superficial accuracy metrics and provide a granular, real-time view of how agents behave in live settings. It doesn't just flag errors; it surfaces the underlying context behind failures, revealing why an agent struggled with a particular user intent or environmental variable. This insight is invaluable for engineering teams, as it transforms evaluation from a passive checkpoint into an active improvement tool. With Cortex, companies can catch degradation early, understand the nuances of edge cases, and feed those learnings directly back into Realm for retraining, creating a virtuous cycle of continuous advancement.
Ultimately, Micro1 is not merely a data lab or an evaluation tool it is a comprehensive intelligence engine that connects the entire lifecycle of model development. From the expert human feedback that sharpens raw capability, to the realistic RL environments that forge robust agentic behavior, to the contextual oversight that ensures reliability in production, Micro1 provides the integrated foundation that AI labs and enterprises desperately need. As the industry races toward truly autonomous systems, Micro1 stands as the critical bridge between laboratory breakthroughs and dependable, real-world impact, ensuring that frontier models do not just perform well on paper, but deliver genuine value in the complex, dynamic world they are meant to serve.