AI Engineer (Agentic Systems / Computer Vision / Industrial AI)

Descrição da Vaga

Requisitos

• Built and shipped production AI systems (3+ years, ideally 5+)
• Delivered agentic systems with orchestration, tool use, memory/state, and human-in-the-loop workflows
• Worked with real-world data at scale (video streams, IoT, sensor data, telemetry)
• Strong backend engineering skills (Python, TypeScript, FastAPI)
• Experience with computer vision (OpenCV, YOLO, segmentation, etc.)
• Familiarity with streaming/data pipelines and real-time systems
• Experience integrating with enterprise platforms (ERP, CMMS, WMS, industrial systems)
• Systems mindset: you think about reliability, cost, failure modes, and scale
• Background in robotics, autonomous systems, or industrial environments
• Experience with knowledge graphs / ontologies
• Built platforms for enterprise customers at scale
• You’ve taken AI systems from idea → production → real-world use
• You’re comfortable working autonomously in high-ownership environments
• You care about execution, not just experimentation
• You’re energized by solving messy, real-world problems
• Work on real-world AI systems, not theoretical models
• High ownership, small team, fast-moving environment
• Direct impact on how AI interacts with physical systems and operations

Responsabilidades

• If you’ve built production AI systems and want to work on problems where data is messy, environments are unpredictable, and the stakes are real, this is that kind of role
• You’ll help design and ship end-to-end AI systems that combine agents, multimodal data, and real-world constraints
• Build agent runtimes and orchestration systems across multiple domains (vision quality, predictive maintenance, operations planning)
• Develop multimodal pipelines (video, time-series, sensor data) using computer vision and signal processing techniques
• Design context systems (ontology, knowledge graph, memory layers) to ground decision-making in real-world environments
• Create decision surfaces: dashboards, alerts, workflows, and audit trails used by operators
• Integrate with enterprise systems (ERP, CMMS, WMS, PLCs) and unify complex schemas
• Own systems from data ingestion → model serving → backend → user interface
• Built multimodal pipelines (vision + structured data)
• Shipped 0 → 1 products, ideally at a VC-backed startup
• Designed systems that handle noise, latency, drift, and imperfect data

Benefícios