Senior Computer Vision Engineer

Original Post

**About Agrex.ai** Agrex.ai is a video analytics company turning enterprise CCTV infrastructure into intelligent, real\-time AI systems. We serve 60\+ enterprise clients across retail, manufacturing, logistics, and BFSI — running across 350K\+ camera channels on edge devices and cloud. We are now building next\-generation GenAI and Vision\-Language Model (VLM) capabilities that let enterprises query their video data in natural language. **The Role** We are looking for a Senior CV/DL Engineer to own our AI infrastructure end\-to\-end — from production inference reliability to new model development to VLM\-powered video search. This is a high\-ownership, high\-impact role that works directly with the CEO and Product leadership. **What You'll Own** * End\-to\-end reliability of the DL inference pipeline across 350K\+ camera channels — targeting 99\.9% uptime * Model optimization for production: ONNX/TensorRT, FP16/INT8 quantization, FPS and latency tuning * New vertical\-specific model development (safety compliance, queue analytics, vehicle intelligence, shelf analytics) * VLM pipeline architecture for natural\-language video search * Technical leadership of the CV team — code reviews, architecture decisions, experiment rigor **Must\-Have** * 4–8 years in CV/DL with at least 2 years in production (not research\-only) * Strong Python (NumPy, OpenCV, PyTorch); Linux, Git; C\+\+ a plus * Deep knowledge of detection, segmentation, and multi\-object tracking architectures * Production model optimization: ONNX, TensorRT, quantization, profiling * Experience with real\-time video pipelines: RTSP, H.264/H.265 * Proven track record of reducing false positives in noisy, real\-world conditions **Good to Have** * NVIDIA DeepStream / GStreamer experience * Exposure to VLMs, LLM\-driven agents, or multimodal AI * MLOps: DVC, W\&B/MLflow, Docker, CI/CD for ML * Domain experience in retail, manufacturing, warehouses, or bank branches **KPIs You'll Drive** System uptime · Alert accuracy · FP rate reduction · Models shipped per quarter · Inference FPS/latency · Support ticket reduction **Why Join** * Real scale: 350K\+ cameras, 60\+ enterprise clients — your models run on real video * Full\-stack AI: hardware, edge, and cloud — you work across the entire stack * Direct access to CEO and product roadmap decisions * Meaningful equity at an early\-stage, fast\-scaling company **Education:** B.Tech / BE / M.Tech / MS in CS, EE, Mathematics, or equivalent practical experience. Research publications are a plus but not required — we value shipped systems. Job Type: Full\-time Pay: ₹1,500,000\.00 \- ₹2,000,000\.00 per year Ability to commute/relocate: * Dwarka, Delhi, Delhi: Reliably commute or planning to relocate before starting work (Required) Experience: * Computer vision: 4 years (Required) Work Location: In person