Engineering Manager, Computer Vision & AI (SiteGuard Squad Lead)

Reports to: VP Engineering

About WakeCap

WakeCap builds the connected-worker platform for large-scale construction and industrial sites. Our hardware (smart helmets, anchors, gateways) feeds a real-time SaaS platform that gives site teams worker location, safety compliance, automated mustering, and productivity analytics. We operate across major projects in the GCC.

About the Role

SiteGuard is WakeCap's computer-vision safety product. It detects PPE non-compliance, hazards, and unsafe behavior from site cameras in real time, running both on the edge (close to the camera) and in the cloud, with consistent detection output either way.

As SiteGuard Squad Lead, you own the whole stack: detection models, camera agents, evaluation and MLOps pipelines, and the product surfaces that turn raw detections into safety actions for HSE managers and site supervisors. You lead a small, senior, cross-disciplinary team with high autonomy and direct product impact.

What You'll Own

- The full lifecycle of SiteGuard's detection capabilities (PPE, hazards, unsafe behavior): problem framing, training, evaluation, production deployment.

- Model strategy across classical detectors (YOLO family), transformer-based detectors (RT-DETR, Co-DETR, GroundingDINO), zero/few-shot approaches (CLIP, DINOv2, SAM 2), and VLMs (Gemini Vision, GPT-4V, Qwen-VL) where they outperform purpose-built detectors.

- Edge and cloud camera agents: optimized inference on constrained hardware (quantization, INT8/FP16, TensorRT, ONNX) and consistent output across deployment types.

- Evaluation and MLOps: a golden labeled dataset as single source of truth, regression gates so nothing ships without a measured quality bar, experiment tracking, model registries, dataset versioning, and drift/cost/accuracy monitoring.

- Hiring, performance, culture, and delivery across the squad.

What We're Looking For

- 10+ years in engineering, with progression from senior IC to leadership; 5+ years directly managing engineers (hiring, performance cycles, difficult conversations).

- Proven CV/ML systems in production, including high-throughput, asynchronous video and media processing.

- Strong Python (async-first: asyncio, FastAPI or equivalent), with reliability and observability built in.

- CV foundations across classical and transformer-based detection and segmentation; video frame handling (OpenCV/FFmpeg); foundation models (CLIP, DINOv2, SAM 2); VLMs for scene understanding.

- Multimodal/VLM APIs (Gemini/Vertex, OpenAI, Anthropic): prompt engineering, JSON-schema-constrained output, caching, per-model tuning.

- Edge inference optimization (ONNX, TensorRT, quantization; Jetson, Hailo, or equivalent).

- MLOps: experiment tracking, model registries, dataset versioning, CI for model evaluation.

- Daily hands-on use of Claude Code and Codex, with grounded awareness of the current AI model landscape.

- Exceptional written and spoken English. This is a hard requirement.

Nice to Have

- Safety, surveillance, or video analytics in industrial or construction settings; OSHA/EHS knowledge.

- Synthetic data generation (ControlNet, Stable Diffusion) for training augmentation.

- LLMOps and observability for model-backed services.

- Agentic frameworks (LangChain, LlamaIndex, AutoGen) applied to safety workflows.

- On-prem or edge deployment under intermittent connectivity.

What We Offer

- Competitive salary, performance bonus, and equity.

- High-autonomy role building a safety AI product from the ground up, your name on the architecture.

- Relocation support for candidates joining from outside KSA.

- Health insurance, annual flights, and standard WakeCap benefits.