Project Post Doc Fellow - Atmospheric Sciences & AI/ML

Project Description

SMART VAYU is a project under the Centre of Excellence – ATMAN (Advanced Technologies for Monitoring Air-quality iNdicators) at the Kotak School of Sustainability, IIT Kanpur. SMART VAYU aims to strengthen air quality management through the deployment of indigenous low-cost air quality sensors, real-time environmental monitoring, data-driven decision support systems, and scientific stakeholder engagement. The project focuses on generating reliable air quality information to support evidence-based policy making, environmental planning, and sustainable urban development. By integrating field monitoring, technology, and research, the project contributes to improving air quality and public health outcomes while strengthening environmental governance.

Postdoc - Atmospheric Science

Position Summary A Project Postdoc Fellow will work on developing and deploying a Machine Learning–based framework for district-wise, real-time PM2.5 prediction and hotspot analysis. The role combines atmospheric/air-quality research, large-scale data processing, ML/deep learning, spatial-temporal analysis, visualization, and coordination with external partners. 

Responsibilities

Process and conduct time-series analysis of air-quality sensor data.

Apply advanced Machine Learning and Deep Learning models.

Handle missing data and generate hourly PM2.5 predictions.

Validate model performance and outputs.

Visualize PM2.5 concentration patterns using heat maps and satellite-based spatial overlays.

Identify pollution hotspots and analyse temporal trends across districts.

Coordinate with external partners for smooth project execution.

Conduct outreach and training in later stages to support policy-makers in using the developed tools. 

Qualifications

Ph.D. in Civil Engineering, Environmental Sciences, Atmospheric Sciences, Climate Sciences, Mathematics, Physics, Climate Policy, or a relevant field.

Strong background in atmospheric modelling and air-quality research.

Significant experience handling and processing large datasets.

Strong publication record in high-impact air-quality research.

Knowledge of scientific programming.

High proficiency in programming languages such as FORTRAN, Python, MATLAB, Igor, R, etc.

____________________________________________________________________________________________________________

Postdoc - AI-ML

Position Summary

A Project Post-Doctoral Fellow will contribute to interdisciplinary research involving Machine Learning, Artificial Intelligence, Deep Learning and statistical/predictive modelling, particularly for large and complex datasets. The role requires strong Python/ML expertise and the ability to collaborate across disciplines.

Key Responsibilities

• Develop and apply ML/AI models for problems involving regression, classification and clustering.
• Work with deep learning and statistical/predictive modelling techniques.
• Analyse large datasets.
• Perform time-series analysis and modelling.
• Use Python and ML/AI libraries such as NumPy, SciPy and scikit-learn.
• Potentially work with deep-learning frameworks such as PyTorch or TensorFlow.
• Explore advanced approaches such as VAR, VARIMA, LSTM, GRU, Transformers and Gaussian Processes.
• Work in an interdisciplinary research environment and incorporate domain knowledge into ML/AI models.
• Work with raw data from instruments or sensors where required.
• Maintain high ethical standards in data analysis and reporting.
• Contribute to high-quality research publications.

Required Qualifications

• Ph.D. in Computer Science, Electrical Engineering, or closely related fields, with specialization in Machine Learning/Artificial Intelligence.
• Strong background in:
• Machine Learning
• Deep Learning
• Statistical/predictive modelling
• Mathematics, particularly probability theory and time-series analysis
• Fluency in Python and relevant ML/AI libraries.
• Significant experience handling and processing large datasets.
• Strong publication record in high-impact ML/AI conferences and/or journals.