ROLE / RESPONSIBILITIES
– Implement and fine-tune modern detection/classification architectures like YOLOv8, Faster/Mask RCNN, EfficientNet, ConvNeXt, and Vision Transformers.
– Conduct ablation studies, hyperparameter tuning, and model optimization experiments (e.g., pruning, quantization, distillation).
– Define labeling guidelines, manage annotation QA loops, and handle class imbalance strategies like re-sampling or focal loss.
– Build data loaders and augmentation pipelines using libraries such as Albumentations or TorchVision, tailored to challenging industrial imagery.
– Design reproducible experiments with clear metric dashboards (mAP, F1 score, PR curves).
– Perform error analysis and model debugging to uncover edge-case failure modes.
– Package models for deployment on cloud services (e.g., Azure, AWS).
– Integrate models into production workflows using REST APIs, Docker, and CI/CD pipelines.
– Work closely with cross-functional teams to translate real-world use cases into model specs.
– Document code and experiments thoroughly and contribute to weekly research reviews.

REQUIREMENTS / QUALIFICATIONS
– 1–3 years hands-on experience in computer vision or deep learning roles.
– Experience with industrial/safety inspection datasets (e.g., PPE detection, visual defect classification).
– Familiarity with MLOps tools like MLflow, DVC, or ClearML.
– Experience with model optimization and deployment frameworks (ONNX, TensorRT, OpenVINO).
– Exposure to real-time or edge inference performance constraints.
– Contributions to open-source, research publications, or competitive CV challenges (e.g., Kaggle).
– Proficiency in Python and deep learning frameworks (PyTorch /Tensorflow).
– Good understanding of CNNs, transfer learning, data augmentation, and overfitting mitigation.
– Familiarity with basic software engineering practices (git, code reviews, unit testing).
– Solid grasp of linear algebra, probability, and optimization as applied in ML.
– Python, PyTorch, Tensorflow, TorchVision, FastAPI, OpenCV
– ONNX, TensorRT, Albumentations
– Docker, Git, Azure/AWS
– Jetson devices, cloud APIs, SQL/NoSQL databases

BENEFITS / PERKS
– Health Insurance
– Flexible Working hours
– 5 Days WFH
– Unlimited ML
– Subsidiary Food options
– Unlimited Beverages