Skip to main content
Association for Academic Excellence

Why we need AI.Research?

AI Research is vital to advance capabilities, solve global challenges, and ensure ethical, safe, and robust systems for the future.

🚀 The Pragyan AI Research Fellowship: Cultivating the Next Generation of AI Scholars

The Pragyan AI Research Fellowship (Pragyan meaning "wisdom" or "supreme knowledge") is a cutting-edge program for CBSE high school students, designed to foster original research and professional application in Artificial Intelligence (AI) and Machine Learning (ML). The Pragyan AI Research Fellowship is an intensive program designed to transform high school students into proactive AI researchers. Fellows gain practical, hands-on project exposure by immediately working on cutting-edge AI challenges. The program provides the mentorship and structured environment necessary to conduct rigorous, independent research. A core benefit is the unique opportunity to develop their work for potential publication in academic journals or showcasing in major science competitions, building a powerful and distinctive portfolio for future academic pursuits.

Get Started Now

Live Lectures

Eight weeks of immersive live training, one 90-minute session weekly, mastering machine and deep learning fundamentals curated by our experts and led by AI research mentors.

  • Regression models & classification algorithms
  • K-means clustering & decision trees
  • Neural networks & convolutional neural networks
  • Hands-on experience building AI/ML projects

Research Publishing

Tackle groundbreaking AI/ML challenges and publish contributions in top-tier academic journals and global conferences.

  • Original research projects
  • Peer review process
  • Publication in top-tier venues

🚀 NeurIPS High School Projects Track🎓

The world's most prestigious AI conference, NeurIPS, has introduced a dedicated high schooler track. This is your chance to present research at the same venue as leading AI researchers from Google, OpenAI, and MIT!

World's Top AI Conference
High School Track Available
Career-Defining Opportunity

The Pragyan AI Research Fellowship, with its focus on practical projects and publication opportunities, is now perfectly aligned to prepare students for submitting to this track.

Prepare for NeurIPS with Pragyan AI Research Fellowship

Content Inspired from the Best Minds

Our program is inspired by the top cited papers on NLP, CV and Data analysis designed and proposed by PhD graduates from world's top institutions and taught by expert AI/ML researchers

Why AI/ML for CBSE High School Students?

AI is transforming our world at an unprecedented pace. Starting research early empowers students to become AI-ready citizens who can shape the future rather than just adapt to it. Early exposure to research builds critical thinking and positions students as innovators in the AI revolution.

Get Started Now

College Application Advantage

Stand out in college applications by conducting high-impact research and gaining exposure through publications and prestigious conference presentations.

Early AI/ML Foundation

Build a strong foundation in AI and machine learning concepts from an early age, giving you a head start in understanding the technologies that will shape the future.

Research Experience

Gain hands-on experience with cutting-edge AI research and develop the critical thinking skills essential for scientific innovation.

Selected Research Spotlight Papers 

Selected Research Spotlight Papers represent high-impact, innovative work from top AI conferences like NeurIPS. They highlight breakthroughs in LLM reasoning, Generative AI efficiency, and crucial advancements in AI safety and robustness, guiding the next wave of fundamental machine learning research.

The Illusion of Thinking: Understanding the Strengths and Limitations of Reasoning Models

This work explores how Large Reasoning Models (LRMs) handle complex tasks, finding that their accuracy collapses beyond certain complexity thresholds. It suggests that while LLMs outperform on simple tasks, LRMs show an advantage only in medium-complexity scenarios, and both fail at high complexity.

The Structure of Relation Decoding Linear Operators in Large Language Models
​

This work Investigates how relational facts are encoded in LLMs. Findings suggest that LLMs encode property-based structure (e.g., country of X) rather than distinct, separate relations, which explains their compressibility and generalization patterns.



STARFlow: Scaling Latent Normalizing Flows for High-resolution Image Synthesis
​

This Work presents a scalable approach for generating comparable quality, high-resolution images using Latent Normalizing Flows, but without the high computational cost and complexity of prior methods like standard diffusion models.




LinEAS: End-to-end Learning of Activation Steering with a Distributional Loss 

A novel approach (LinEAS) for mitigating issues like toxicity in language models by learning to directly steer the model's internal activations with minimal unpaired samples, making interventions precise and effective.

ModHiFi: Identifying High Fidelity predictive components for Model Modification

This work proposes a method for model modification (e.g., pruning or unlearning) without needing access to the original training data or loss function, relying instead on identifying critical components via a metric called Subset Fidelity.

Agnostic Active Learning Is Always Better Than Passive Learning
​​

A theoretical paper that provides strong characterization of the complexity and superiority of agnostic active learning over traditional passive learning across various concept classes.


About AAE

About Mentor

MENTOR
Subroto Chowdhury, Ph.D., CAIIB

Subroto Chowdhury is a Practitioner Professor, Researcher, and Independent Consultant specializing in the practical application of Artificial Intelligence and Statistical Modeling.

His academic background includes a Ph.D. for thesis in Econometrics and Statistical Modeling. His career is marked by a unique blend of industry and research roles, including banking at Bank of Baroda, research at NIBM Pune (an RBI institute), and teaching at NIT Silchar.

A certified expert in the field, he holds exposure to Artificial Intelligence from the University of Helsinki, along with expertise and certifications in Python Programming and Machine Learning from industry giants IBM and Google. He uses his diverse experience and certified knowledge (CAIIB) to deliver high-impact lectures, transforming complex concepts into accessible, practical wisdom.


Over 20+ Years of Industry Experience in Banking and Finance Domain
Formerly Associated with Bank of Baroda, NIT Silchar, NIBM Pune, HDFC Bank and Many Others
2022 Award Winner RBI Innovation Hub
2024 Award Winner Best Case Study, CIMP,Patna

INTERACTIONS

Lecture, Talks, Presentations, Awards

Class in Progress @ NIT Silchar
​B.Tech Students

Workshop on Statistical Modeling in R
​@National Institute of Technology Silchar

2020 Artificial Intelligence Exposure @University of Helsinki

Class in Progress @ NIT Silchar
​MBA Students

Presentation in International Conference @ IIT Kharagpur

Talk in Progress @ NIT Silchar 2018

Accomplishments

15+

Recognitions

20+

Cases and Papers 

150+

Speeches

4k+

Student Connect

05+

Awards

Recognition

 Excellence, Integrity, and Distinctions

FEATURED WORK

Workshop and Hands on Training on Python ,R, EVIEWS

Get More
GET IN TOUCH

Contact us for any enquiry.

Address

SS 22, Financial Skills Hub, Durgapur - 713213

Phone

+91- 6294250640

Hours

10 am to 7 pm 
Monday to Friday

Email

contact@aeindia.org