Job Description

Location: Jaipur/Remote (Candidates from any location can work for this position remotely, but they will be required to visit the Jaipur headquarters once every month)

• Experience with computer vision pipelines such as object detection (YOLO family), tracking (DeepSORT/ByteTrack), or marker detection (ArUco/AprilTags).
• Strong understanding of PyTorch and at least one modern detection framework (Ultralytics, SuperGradients, MMDetection).
• Familiarity with multi-frame temporal reasoning, identity tracking, and feature embedding–based matching.
• Exposure to GPU optimization on NVIDIA hardware (TensorRT, CUDA basics, mixed precision, batching strategies).
• Experience with distributed processing frameworks like Ray (preferred) or Dask for scaling inference workloads.
• Hands-on work with real-time video ingestion/streaming (RTSP, Kafka/MQTT, async queues).
• Understanding of ReID models, cosine similarity scoring, and association logic.
• Ability to diagnose occlusion, blur, and orientation-related detection issues and tune confidence thresholds.
• Experience deploying CV services in Docker, ideally with Ray Serve or FastAPI-based microservices.
• Familiarity with relational databases (PostgreSQL/MySQL) for storing and validating event data.