About Us

We are establishing the first distributed Al infrastructure dedicated to personalized Al. The evolving needs of a data-driven society are demanding scalability and flexibility. We believe that the future of Al is distributed and enables real-time data processing at the edge, closer to where data is generated. We are building a future where a company's data and IP remains private and it's possible to bring large models directly to consumer hardware without removing information from the model.

Role Overview

As a Staff R&D AI Engineer, you will lead the development of cutting-edge AI systems that bridge computer vision, natural language understanding, and action learning. You'll architect and implement Vision-Language-Action (VLA) models, advance reinforcement learning applications, and push the boundaries of multimodal AI integration. This role combines deep expertise in both computer vision and large language models with hands-on experience in reinforcement learning to create intelligent systems that can understand, reason about, and interact with complex environments. You'll drive research initiatives, mentor technical teams, and translate breakthrough AI research into practical applications across diverse domains.

Key Responsibilities

Design and develop Vision-Language-Action (VLA) models that integrate visual perception, natural language understanding, and action prediction

Architect and implement reinforcement learning systems for sequential decision-making, including policy learning and skill acquisition

Build and optimize computer vision pipelines for perception tasks, including object detection, segmentation, tracking, and scene understanding

Develop and fine-tune large language models for instruction following, reasoning, and task planning applications

Implement RLHF (Reinforcement Learning from Human Feedback) systems to improve model alignment and safety

Create multimodal training pipelines that leverage synthetic and real-world data for robust model performance

Research and prototype novel AI architectures that combine vision, language, and action learning

Collaborate with engineering teams to integrate AI models into applications and validate performance across domains

Optimize model inference performance for real-time applications across edge and cloud deployments

Lead technical initiatives, mentor junior AI engineers, and establish best practices for AI model development

Stay current with latest research in VLA models, multimodal AI, and robotics to drive innovation roadmap

Present findings at conferences and publish research to advance the field

Qualifications & Skills

7+ years of experience in AI/ML engineering with 4+ years focusing on deep learning and neural network development

Strong understanding of reinforcement learning algorithms and their applications (PPO, SAC, TD3, etc.)

Strong expertise in both computer vision and natural language processing with hands-on model development experience

Proficiency in PyTorch and/or TensorFlow with experience training and deploying large-scale models

Experience with transformer architectures, attention mechanisms, and large language model fine-tuning

Hands-on experience with computer vision tasks including object detection, semantic segmentation, and visual tracking

Strong programming skills in Python with experience in distributed training and model optimization

Understanding of sequential decision-making and control systems fundamentals

Experience with MLOps practices including model versioning, monitoring, and deployment pipelines

Proven ability to work independently on complex research problems and deliver practical solutions

Strong communication skills and experience collaborating with cross-functional engineering teams

Preferred Qualifications

PhD in Computer Science, Robotics, AI/ML, or related field with focus on multimodal learning or robotics

Direct experience developing or working with Vision-Language-Action (VLA) models or similar multimodal architectures

Experience with RLHF implementation and human feedback integration for model alignment