Job Description:
- Development, adaptation, and implementation of AI/ML algorithms and frameworks, Prediction algorithms
- Developing deep learning and machine learning algorithms (CNN, object detection, segmentation, SVM, AE)
- Time series forecasting: AR, ARIMA, SARIMA, ES, Prophet, LSTM
- Conduct data preprocessing, augmentation, and annotation workflows for image datasets.
- Design, train, and validate deep learning architectures for feature identification using CNN, ResNet,
- EfficientNet, YOLO, U-Net, Mask R-CNN, ViT/Swin Transformer.
- Develop clean, modular, and production-ready code for model training, inference, and deployment.
- Collaborate with domain experts to translate agricultural knowledge into AI models.
- Support integration of models with mobile application (through APIs and deployment-ready formats like TensorFlow Lite / ONNX).
- Write unit tests, integration tests, and documentation to support long-term use of the framework.
- Document methodologies, benchmarking reports, and prepare technical handover materials.
Minimum Qualifications and Experience:
- B.Tech in Computer Science, Electronics and Communications or any related field with 3-5 years of relevant experience
OR
- M.Tech in Computer Science, Electronics and Communications or any related field with 2-3 years of relevant experience
Required Expertise:
- Strong hands-on experience with Python and ML/DL frameworks (PyTorch, TensorFlow, Keras).
- Proficiency in computer vision techniques – CNNs, object detection (YOLO/SSD), segmentation (U-Net/Mask R-CNN), Vision Transformers (ViT, Swin Transformer, DeiT).
- Libraries: NumPy, Pandas, OpenCV, Scikit-learn, Matplotlib/Seaborn.
- Knowledge of model optimization for deployment (quantization, pruning, TensorFlow Lite, ONNX).
- Experience in developing APIs (Flask/FastAPI) for model serving.
- Familiarity with ETL processes, data pipelines, and statistical validation methods.
- Basic understanding of Docker and version control (Git) and experience with MLOps tools
- Ability to write production-grade Python code following best practices (modular design, logging, testing, error handling)
- Understanding of statistical analysis such as normality test, dicky fuller test etc
Location of work:
- TIH-IoT, IIT Bombay Campus, Powai, Mumbai 400076.