Computer Vision (Embedded AI – Dashcam Perception)

About the internship: Selected intern's day-to-day responsibilities include: 1. Train, fine-tune, and evaluate object detection, keypoint estimation, and classification models for road-facing and cabin-facing AI dashcam applications using PyTorch/TensorFlow. 2. Build, maintain and version large-scale image and video datasets from production fleet footage, including annotation, quality control, hard-negative mining and reproducible train/validation/test splits using CVAT or Roboflow. 3. Analyse model failures under real-world Indian driving conditions including low-light, IR night vision, rain, glare, dust, occlusions and dense mixed traffic, and translate observations into targeted data collection strategies. 4. Benchmark model performance using mAP, IoU, Precision, Recall, inference latency, FPS, memory consumption and NPU utilisation while balancing accuracy and deployment constraints. 5. Optimise deep learning models for edge deployment through INT8 quantization, quantization-aware training, pruning, ONNX/TFLite conversion and benchmarking on ARM + NPU hardware. 6. Work alongside firmware engineers to evaluate camera pipeline parameters including ISP tuning, exposure, white balance, WDR behaviour and IR-cut switching, and study their impact on perception accuracy. 7. Assist in multi-sensor perception experiments involving camera, IMU and LiDAR data for event detection and validation. 8. Maintain experiment logs, reproducible training pipelines, model documentation, configuration management, and version-controlled code repositories. Who can apply: Only those candidates can apply who: are available for full time (in-office) internship can start the internship between 24th Jul'26 and 28th Aug'26 are available for duration of 6 months have relevant skills and interests are Computer Science Engineering students * Freshers looking for jobs may also apply Stipend: INR₹ 10,000 - 15,000 /month Deadline: 2026-08-23 23:59:59 Other perks: Certificate, Letter of recommendation, Informal dress code, 5 days a week, Job offer Skills required: OpenCV, Python, Linux, Git, Machine Learning, Computer Vision, Deep Learning, TensorFlow, PyTorch, NumPy and Yolo Other Requirements: 1. Pursuing or recently completed B.E./B.Tech/M.Tech/M.Sc. in Computer Science, Artificial Intelligence, Electronics & Communication, Electrical Engineering, Data Science or a related discipline. 2. Strong hands-on experience with Python, NumPy, OpenCV and at least one of PyTorch or TensorFlow through research, internships, academic projects, Kaggle competitions or personal work. GitHub or portfolio links are strongly preferred. 3. Candidates should possess a solid understanding of modern object detection concepts including anchor-based vs anchor-free detectors, NMS, IoU thresholds, confidence calibration, mAP evaluation metrics, dataset preparation and Linux/Git workflows. 4. Exposure to YOLO, ONNX, TensorRT, OpenVINO, TFLite, CVAT, Roboflow, MMDetection, Albumentations, MediaPipe, Jetson, Raspberry Pi or other embedded AI deployment platforms will be considered an added advantage About Company: We are an Ahmedabad-based AI tech company specializing in dashcams. Additionally, we are the Indian subsidiary of an already established US firm. Additional information: Job offer: On successful conversion to a permanent employee, the candidate can expect a salary of ₹ 400000 to 600000/year