Artificial Intelligence Researcher
EllanorAI
Posted 18 Nov 2026, 11:49 IST
This job at a glance
- Who can apply
- Freshers and up to 1 year · M.Tech or M.Sc
- Must know
- StatisticsPyTorchPython
- Nice to have
- Deep learning
- Work
- On-site in Mumbai
- Pay
- Not stated by the employer.
- Skill demand
- Python70%Statistics27%PyTorch24%Deep learning21%of 99 similar fresher jobs posted in India this week
- Listed
- 18 Nov 2026 · still open
Sign in free to see which of these skills your resume already shows.
About the Role
EllanorAI is a deep-tech AI research startup developing novel architectures for scalable, efficient, and adaptive artificial intelligence. We are seeking an AI Researcher to join our core research initiative, Adaptive Scale Networks (ASN). This is a research-intensive role for someone who wants to work beyond conventional Transformer architectures and contribute to the development of new approaches to foundational AI systems.
Key Responsibilities
- Conducting theoretical and experimental research on Adaptive Scale Networks (ASN).
- Developing novel neural network architectures, adaptive computation mechanisms, and dynamic routing systems.
- Researching spectral methods, signal processing, linear algebra, and their applications to deep learning.
- Designing and implementing efficient attention mechanisms and feed-forward network alternatives.
- Investigating computational complexity, parameter efficiency, memory consumption, and training scalability.
- Developing and evaluating novel approaches to language-model pretraining, fine-tuning, and inference.
- Designing rigorous experiments, benchmarks, ablation studies, and statistical evaluations.
- Translating mathematical formulations and published research into functional PyTorch implementations.
- Analysing experimental results and contributing to research papers for peer-reviewed academic publications.
- Collaborating directly with the founders on architectural research and long-term research direction.
Requirements & Tech Stack
Minimum Qualifications:
- M.Tech / M.Sc. in Computer Science, Applied Mathematics, Mathematics, Computational Neuroscience, or a closely related discipline.
- Strong foundations in linear algebra, calculus, probability, statistics, and mathematical modelling.
- Practical experience with Python and PyTorch.
- Good understanding of deep learning architectures, particularly Transformers and modern language models.
- Ability to independently read, understand, and implement research papers.
- Experience designing experiments, analysing results, and developing reproducible research workflows.
Preferred Qualifications:
- PhD completed or currently pursuing a PhD in Computer Science, Mathematics, Neuroscience, or a closely related research discipline.
- Research experience in neural architecture design, efficient deep learning, reinforcement learning, computational mathematics, or foundation models.
- Publications in reputed peer-reviewed conferences or journals (NeurIPS, ICML, ICLR, AAAI, JMLR, SIAM, or related venues).
- Experience with spectral methods, matrix decomposition, SVD, orthogonal transformations, numerical optimisation, or dynamical systems.
- Experience with distributed training, GPU optimisation, CUDA, Triton, or high-performance computing.
What Will Give You an Advantage:
- Previous research on alternatives to conventional Transformer architectures.
- Experience with adaptive neural networks, conditional computation, mixture-of-experts, or dynamic routing.
- Knowledge of signal processing, Fourier analysis, wavelets, or frequency-domain representations.
- Experience building and training language models from scratch.
- Strong mathematical research ability and a track record of independent scientific contributions.
What You'll Get
- Direct involvement in developing a novel foundation-model architecture.
- Opportunity to contribute to original AI research rather than simply integrating existing models.
- Close research collaboration with the founder and core engineering team.
- Access to real-world model development, experimentation, and GPU-based research infrastructure.
- Opportunities to co-author research publications and contribute to intellectual property.
- The opportunity to help shape the technical direction of a long-term AI research initiative.
Who Should Apply?
We are looking for someone with a genuine research mindset: someone comfortable with mathematical depth, experimental uncertainty, and challenging established approaches. If you are pursuing or have completed a PhD and want to contribute to the next generation of neural network architectures, we would like to hear from you.
Apply: Send your CV, research profile, Google Scholar / publication links (if available), and a brief statement of your research interests to [email protected].