Technical Architect

overview one - you’ll need to know, in practice, how model serving actually works at scale, where inference bottlenecks show up, and how the levers between accuracy, latency and cost actually trade off against each other. You should also be comfortable enough with computer vision (CNNs, vision transformers, detection and segmentation methods) to push back on a data science team’s approach when needed. Day to day, you’d be:Taking ownership of the model lifecycle - pipelines, registry, versioning, CI/CD, drift detection, and retraining logicWorking out how to run models in constrained or edge settings: quantisation, pruning, distillation, ONNX/TensorRT, and picking the right accelerators for the jobSetting the evaluation approach - where precision and recall trade off, where thresholds sit, and how much false-positive friction is acceptable for the people using the systemDesigning so the model supports human decision-making rather than overrides itBuilding for sites with poor or intermittent connectivity, with training centralised and models pushed out from thereGetting the non-functional side right at scale - throughput, latency, availability, resilience, DR - across multiple sitesReviewing and challenging supplier architecture, including calling out vendor lock-in risk and build-vs-buy callsWorking to Secure by Design principles and NCSC guidance throughout, given the OFFICIAL-SENSITIVE classification What you’ll bringA track record as a solution/technical architect on AI projects sitting inside bigger strategic programmesSolid AWS experience - S3, Lambda, EventBridge, SNS/SQSComfortable working across TOGAF, ArchiMate, C4, UML and BPMNHands-on with MLOps tooling and platforms - SageMaker, OpenVINO or equivalentExperience architecting data for large volumes of imagery - tiering, retention, lineage, provenanceFamiliar with the governance side of AI - DPIAs, model documentation, bias and fairness checksStrong stakeholder handling, and the ability to explain architecture to people who aren’t architectsThis role requires active SC Clearance Bonus points forBackground in X-ray-based AI models or density/object identificationPublic sector delivery experience - GDS standards, Technology Code of Practice, spend controlsKnowledge of the DSIT AI Playbook or the Algorithmic Transparency Recording StandardFinOps experience specifically around GPU and inference costExperience integrating with scanner/hardware OEM systems and real-time image pipelines