Summary description:

Spiral Networks is looking for a mission-driven AI Systems Architect to join the team — our real-time AI-powered situational awareness platform. In this role you will design, build, and stress-test the AI pipeline that processes live video feeds, detects anomalies, and flags threats before they escalate. You will own the architecture from sensor input to alert output, while simultaneously acting as the team's red team lead — actively trying to break, fool, and bypass the system so we can harden it for real-world deployment. This is a builder's role at the intersection of computer vision, security engineering, and adversarial machine learning.

Job description:

Design the video AI pipeline. Architect end-to-end ingestion, pre-processing, and inference workflows for live and recorded video streams using cloud-native and edge infrastructure.Build and tune detection models. Train, fine-tune, and evaluate computer vision and anomaly detection models. Establish benchmarks for accuracy, latency, and false-positive rates that meet operational thresholds.Conduct adversarial testing. Execute structured red-team campaigns against the AI system — crafting evasion attacks, data poisoning scenarios, environmental spoofing, and sensor manipulation tests to expose and close vulnerabilities before deployment.Document and brief findings. Produce clear technical reports, threat models, and risk assessments. Present findings to engineering and leadership in structured after-action style reviews.Mentor through live OJT. Work side-by-side with senior architects and SkillBridge participants in synchronous code walk through, sessions, meetings, and structured check-ins — minimum 120 hours of hands-on collaborative training across the three-phase program.Technical skills we are looking for:Computer vision & video AIAdversarial ML / red teamingPython, PyTorch, or TensorFlowAWS or cloud GPU infrastructureNVIDIA CUDA / edge inferenceSystems architecture & APIsCI/CD, MLOps, model monitoringSecurity frameworks & threat modeling

Eligibility factors: