MLOps Engineer - Senior Level

**We’re Hiring: Senior Computer Vision / Machine Learning Engineer** **AI Video Analytics \| Retail Loss Prevention \| Cashier Monitoring** ExcelsTech/Univenture is building an AI\-powered cashier monitoring and theft\-prevention system for retail environments. This is not a generic AI project. This is not a simple CCTV analytics task. This is a real\-world computer vision product where the AI must understand cashier\-area video, hand movement, cash drawer activity, suspicious behavior patterns, event clips, confidence scoring, and human review workflows. We are looking for a **Senior Computer Vision / ML Engineer** who can help us build the core AI pipeline from video footage to usable event evidence. What you will work on: You will help design and build AI models and video analytics workflows for: ✅ Hand detection and hand movement tracking ✅ Cash drawer zone monitoring ✅ Cash and object detection ✅ Suspicious hand movement near cash drawers ✅ Hand lingering / dwell\-time detection ✅ Cash movement away from drawer or counter zones ✅ Event classification and confidence scoring ✅ Video clip generation for review ✅ False\-positive / false\-negative analysis ✅ AI model evaluation and improvement ✅ Future edge AI deployment readiness The system will support a human review workflow. AI output will be used as **decision\-support evidence** , not as an automatic accusation or final judgment. Required skills We are looking for strong practical experience in: * Python * PyTorch or TensorFlow * OpenCV * YOLO\-family models or equivalent object detectors * Video analytics and frame\-by\-frame processing * Object detection and custom model training * Multi\-object tracking such as ByteTrack, DeepSORT, StrongSORT, BoT\-SORT, or similar * Hand/keypoint tracking such as MediaPipe Hands or custom pose/keypoint models * ROI / zone\-based event logic * Temporal behavior detection over short video windows * Dataset preparation and annotation workflows * Model evaluation using precision, recall, false positives, and false negatives Nice to have Experience with any of the following will be a strong plus: * NVIDIA Jetson / TensorRT / DeepStream * RTSP / ONVIF camera streams * Edge AI deployment * AWS / GCP / cloud\-hosted AI workflows * FastAPI / Flask for model APIs * CVAT, Label Studio, Roboflow, or similar annotation tools * Retail analytics, surveillance AI, or loss\-prevention systems * POS / transaction log correlation * MLOps tools such as MLflow, W\&B, or DVC Who we are looking for This role is ideal for someone who can build practical AI systems, not just train models in notebooks. You should be comfortable working with real\-world video challenges such as: * Poor camera angles * Changing lighting * Occlusion * Hand overlap * Low\-quality footage * Limited suspicious\-event examples * False\-alarm control * Edge deployment constraints We value engineers who are honest about model limitations, careful with evaluation, and strong at turning messy data into working systems. Location 📍 Lahore, Pakistan preferred 🌍 Hybrid / remote considered for exceptional candidates How to apply Please send: * Your CV * GitHub / portfolio / project examples * Any relevant work in computer vision, video analytics, object detection, tracking, action recognition, or edge AI Interested candidates can share their CV's at seerat@excels\-tech.com or WhatsApp at 03330555321\.