Project Research Scientist-I

Indraprastha Institute of Information Technology
West District, Delhi

Apply Now

Job details

₹56 an hour
6 days ago

Qualifications

TensorFlow
Computer science
MATLAB
Signal processing
Git
Master's degree
Analysis skills
Math
Docker
Bachelor's degree
Machine learning
Back-end development
REST
Django
Software development
APIs
Software testing
Data science
Flask
AI
Communication skills
Python

Full job description

Subject: Applications are invited for three Project Research Scientist–I (Non-Medical) positions at
SBILab and HMI Lab of Indraprastha Institute of Information Technology Delhi (IIIT-Delhi)..

The selected candidates will contribute to an interdisciplinary healthcare research project involving
multimodal clinical data, electrophysiological signals, neuroimaging, artificial intelligence, patient-specific
brain simulations, and development of a clinically deployable decision-support platform.

Lab Website: https://sbilab.iiitd.edu.in; https://hmi.iiitd.edu.in/

Post and Position:

Project Research Scientist - I: 03 Positions

The three positions will have complementary technical responsibilities:

1. Multimodal AI for seizure analysis
2. Simulation-informed AI and TVB integration
3. MANAS platform development, MLOps, and deployment

Reporting to:

Prof. Anubha Gupta
Signal Processing and Biomedical Imaging Laboratory (SBILab),
Indraprastha Institute of Information Technology Delhi (IIIT-Delhi)

Dr. Jainendra Shukla
Associate Professor,

Human-Machine Interaction Lab (HMI Lab)
Indraprastha Institute of Information Technology Delhi

The researchers will also interact closely with clinical and neuroscience collaborators from AIIMS, New
Delhi involved in multimodal data acquisition, annotation, brain-network modelling, clinical validation,
and presurgical epilepsy evaluation.

Fellowship:

Rs. 56,000/- + HRA (as admissible) = approx. Rs. 72,800/- inclusive

Essential Qualifications:

The candidate should possess:

Post Graduate Degree, including the integrated PG degrees.

OR
For Engineering / IT / CS - Four Years Graduate Degree.

Require strong mathematical and coding skills

Relevant disciplines include Computer Science, Data Science, Artificial Intelligence, Electronics,
Mathematics, or allied areas.

Candidates should also possess:

Strong mathematical, analytical, and programming skills
Proficiency in Python and familiarity with machine-learning libraries
Ability to work with large and heterogeneous datasets
Evidence of research aptitude through projects, publications, thesis work, software

development, or relevant professional experience

Strong written and oral communication skills
Ability to work effectively in an interdisciplinary team involving engineers, neuroscientists,

and clinicians

Position 1: Multimodal AI for Seizure Detection, Classification, and Prediction

Primary Responsibilities: The selected candidate will:

Develop deep-learning pipelines for analysis of electrophysiological data, particularly EEG

and MEG.

Develop models for seizure detection, seizure classification, and short-horizon seizure

prediction.

Extract and model temporal, spectral, spatial, and connectivity-related features from

electrophysiological signals.

Integrate electrophysiological features with neuroimaging-derived and structured clinical

variables.

Develop and compare temporal architectures such as convolutional neural networks, temporal

convolutional networks, recurrent neural networks, and transformer-based models.

Address class imbalance, inter-patient variability, calibration, and uncertainty estimation.
Establish patient-wise training, validation, and testing pipelines that prevent data leakage.
Benchmark developed models against conventional machine-learning and deep-learning

baselines.

Evaluate models using clinically relevant metrics such as AUROC, precision-recall AUC,

sensitivity, specificity, false-alarm rate, calibration, and prediction lead time.

Contribute to publications, technical documentation, intellectual-property generation, and

project reporting.

Preferred Expertise

Applicants with experience in one or more of the following will be preferred:

EEG, MEG, biomedical time-series, or physiological signal processing
PyTorch, TensorFlow, or JAX
Transformer, CNN, LSTM/GRU, or temporal convolutional architectures
Self-supervised or foundation-style representation learning for time-series data
Signal preprocessing using Python, MATLAB, MNE-Python, EEGLAB, or related tools
Statistical validation of biomedical AI models

Position 2: Simulation-Informed AI and Multimodal Epileptogenic-Zone Localisation

Primary Responsibilities: The selected candidate will:

Develop AI models for direct localisation of epileptogenic zones from multimodal clinical

data.

Construct region-wise multimodal representations using imaging, electrophysiological,

metabolic, and clinical features supplied through the project.

Integrate patient-specific outputs generated using The Virtual Brain (TVB) with data-driven

AI models.

Develop methods for comparing independently generated AI-based and simulation-based

localisation results.

Perform concordance and discordance analyses between direct AI predictions,

connectome/TVB-based inference, SEEG findings, and surgical outcomes.

Develop graph-based and multimodal-learning models where graph structures or connectomic

representations are available from collaborating partners.

Evaluate graph-convolutional, graph-attention, multimodal-fusion, and interpretable ensemble

approaches.

Generate region-wise epileptogenicity probability maps with confidence and uncertainty

estimates.

Assess localisation using Dice similarity, regional sensitivity and specificity, Cohen’s kappa,

SEEG concordance, and association with surgical outcomes.

Develop confidence-aware methods for integrating concordant AI and simulation outputs.
Contribute to publications, intellectual property, reproducible pipelines, and project reports.

Preferred Expertise

Applicants with experience in one or more of the following will be preferred:

Medical image analysis or multimodal biomedical data analysis
Graph neural networks, graph representation learning, or connectomic analysis
MRI, DTI, fMRI, PET, SPECT, EEG, or MEG-derived features
Multimodal fusion, domain adaptation, or missing-modality learning
Model calibration, uncertainty estimation, and interpretable machine learning
Neuroinformatics, computational neuroscience, or patient-specific simulation frameworks
Familiarity with TVB or comparable brain-network simulation environments will be

advantageous but is not mandatory

Position 3: MANAS Platform, MLOps, and Clinical AI Deployment

Primary Responsibilities: The selected candidate will:

Develop and integrate AI components into the proposed MANAS—Multimodal AI-based

Neuro Analysis Software platform.

Design reproducible model-training, evaluation, packaging, and inference workflows.
Develop backend services and application programming interfaces for AI and simulation

modules.

Implement model versioning, experiment tracking, audit trails, and reproducible deployment

pipelines.

Integrate outputs from seizure-analysis models, localisation models, and TVB-based

simulations into a unified decision-support workflow.

Develop clinician-facing visualisations, including region-wise probability maps, seizure-risk

estimates, confidence scores, and evidence summaries.

Support secure, modular, and scalable deployment within institutional computing

environments.

Implement monitoring for model performance, data drift, inference failures, and controlled

model updates.

Optimise models for reliable and latency-aware inference.
Participate in software requirements analysis, modular design, implementation, testing,

documentation, and deployment following software-development life-cycle principles.

Contribute to prospective workflow evaluation, intellectual-property generation, software

documentation, and technology-transfer activities.

Preferred Expertise Applicants with experience in one or more of the following will be preferred:

Machine-learning engineering, MLOps, or AI systems development
Python-based backend development using FastAPI, Flask, Django, or related frameworks
Model deployment using Docker, containers, REST APIs, or workflow orchestration tools
Experiment tracking and versioning using MLflow, Weights & Biases, DVC, Git, or

comparable platforms

Databases, secure data pipelines, dashboard development, or clinical software integration
PyTorch or TensorFlow model packaging and inference optimisation
Software testing, documentation, cybersecurity, or medical-AI deployment

Last Date for Application: 4th Sep, 2026 or until suitable candidate is found.

Duration: Initially for 11 months, subject to successful completion of a three month probation period.
The duration may be extended further depending on performance.

Application Process

Interested candidates are invited to apply through the Google Form:

https://forms.gle/4RLpNWreAkiH15r89

Applicants should clearly indicate their preferred profile:

Profile 1: Multimodal AI and seizure modelling
Profile 2: Simulation-informed AI and EZ localisation
Profile 3: MANAS, MLOps, and deployment

Candidates may indicate more than one profile in order of preference.

The selection committee may decide shortlisting criteria and only shortlisted candidates will be notified via
email and followed up with the interview round.
IIIT-Delhi norms will be followed in hiring and the Institute’s decision will be final in all respects.

Apply Now

Jobseeker tools

Employer Tools

Browse

Stay Connected