Senior Computer Vision Engineer

🚀 Computer Vision Engineer

🌍 Fully Remote | Global Opportunity

🌟 About the Role

We’re partnering with an ambitious, fast-growing company building AI-native workflows for property maintenance and compliance. This isn’t just another AI role this is a chance to apply cutting-edge computer vision and Visual Language Models (VLMs) to real-world problems that directly impact safety, cost, and decision-making.

You’ll be working across high-impact systems where precision, explainability, and reliability are critical, helping shape how AI is applied in a regulated, evidence-driven environment.

🧠 What You’ll Be Doing

🔍 Visual Language Models for Property Diagnostics

• Build and optimise VLM-powered pipelines for diagnostics and guided troubleshooting
• Combine model outputs with structured reasoning frameworks (e.g., severity grading, allowed recommendations)
• Design auditable prompts, inputs, and output schemas that support testing and compliance use cases

⚖️ Severity Grading & Explainability

• Implement autonomous severity grading systems with calibrated confidence thresholds
• Define clear escalation rules for high-risk scenarios
• Collaborate with product and operations teams to ensure outputs are transparent and explainable (evidence used, model uncertainty, etc.)

📱 Edge Computer Vision (Quality & Liability Reduction)

• Deploy lightweight CV models on edge devices (mobile browsers) to deliver real-time feedback
• Enable capture quality checks such as framing, focus, and exposure
• Build “retake required” detection systems to improve data quality upstream
• Optimise for performance, battery efficiency, and device variability, while minimising on-device processing for privacy

🛠️ CV Anti-Fraud & Parts Identification

• Develop contractor-facing CV pipelines to identify parts and flag suspicious claims
• Integrate with supplier APIs for real-time availability and pricing validation
• Design anti-fraud heuristics combining visual signals with structured data

⚙️ AI-First Engineering & Quality

• Leverage AI tooling to accelerate prototyping, training, and failure analysis
• Maintain a high bar for production systems through:
• Dataset versioning
• Reproducible training and inference pipelines
• Offline evaluations and staging tests
• Tight feedback loops from real-world outcomes

✅ Must-Have Skills

• Strong production experience with PyTorch or TensorFlow
• Expertise in the HuggingFace ecosystem (fine-tuning, inference, deployment)
• Hands-on experience with VLMs (e.g., GPT-4V, Gemini Vision, LLaVA, or similar)
• Ability to evaluate API-based vs. locally fine-tuned model approaches
• Experience with ONNX and/or TFLite for edge deployment
• Proficiency in OpenCV (or similar) for preprocessing and validation
• Strong evaluation mindset: metrics, datasets, thresholds, regression prevention

⭐ Nice to Have

• Experience with property diagnostics (e.g., damp/mould detection, thermal imaging)
• Mobile ML experience (Core ML, TFLite)
• Synthetic dataset generation for rare or edge-case scenarios
• Experience in regulated or compliance-heavy environments
• Familiarity with evidence-grade AI systems and outputs

🌍 Why Join?

• Fully remote, globally distributed team
• Work on real-world, high-impact AI systems, not just prototypes
• Tackle meaningful challenges across computer vision, reasoning systems, and applied AI
• Shape how AI is used in safety-critical and compliance-driven industries

👉 Interested? Apply now or reach out to learn more.