Agnideep “Agni” Aich

PhD Candidate (ABD)

“To truly laugh, you must be able to take your pain, and play with it.” – Charlie Chaplin

About Me

Agni's Photo

Hey there! I’m Agni, a Ph.D. candidate in Statistical Machine Learning at the University of Louisiana at Lafayette, working under the guidance of Prof. Bruce Wade. My research explores Statistical Machine Learning, Deep Learning, Applied Machine Learning, Biostatistics, Copula Theory, and Sports Analytics.

But there’s more to me than equations and models. I’m passionate about cars, sunglasses, colognes, soccer, MMA, and EDM — especially trance, future bass, and synthwave. Whether I’m cruising in my Mercedes, producing beats in FL Studio, or chasing the perfect scent, I live for the little things that make life unforgettable.

In everything I do, whether solving a complex statistical problem or crafting an original sound, I bring intensity, curiosity, and creativity. I’m not just building the future, I’m living it.

Research Interests

Statistical Machine Learning

Exploring the theoretical foundations of learning algorithms, with a focus on minimax rates, distribution-free methods, and dependence modeling. I’m particularly interested in how rigorous statistical theory can inform and strengthen modern machine learning systems.

Deep Learning

Designing and analyzing deep neural networks to tackle non-convex optimization challenges, build expressive generative models, and enhance learning under uncertainty. My work combines mathematical insight with practical neural architecture innovation.

Applied Machine Learning

Bridging the gap between theory and application by using copula theory and deploying machine learning techniques to real-world problems, from healthcare and finance to high-stakes decision-making scenarios. Emphasis on interpretability, robustness, and domain relevance.

Biostatistics

Investigating learning frameworks for complex biological datasets, particularly in genomics. My research focuses on feature selection and attention-based neural methods for identifying meaningful patterns in high-dimensional biomedical data.

Sports Analytics

Applying statistical and machine learning models to analyze in-game performance and psychological pressure in elite sports. Current projects use play-by-play data to identify clutch moments and uncover latent performance patterns in competitive environments.

Work Experience

Statistical and Data Science Research Assistant

University of Louisiana at Lafayette
August 2025 - May 2026

I analyze public health datasets using statistical modeling, machine learning, and data visualization under Prof. Amanda Mayeaux. I also collaborate across public health, education, and kinesiology to frame research questions and design analyses, contributing to manuscripts in preparation.

Research Assistant

Statistical Consulting Center, University of Louisiana at Lafayette
August 2024 - May 2025

Provided statistical consulting, data analysis, and data visualization services to PhD students, faculty, and other clients. Also, closely collaborated with clients/researchers to translate objectives into statistical questions and develop robust, actionable solutions.

Instructor of Record

University of Louisiana at Lafayette
August 2020 - May 2024

Taught two sections of Elementary Statistics (Stat 214)/Applied College Algebra (Math 105) each semester. Developed course materials, including lectures, assignments, and exams. Received positive feedback for effective teaching methods and student engagement.

Education

Ph.D. in Statistical Machine Learning

University of Louisiana at Lafayette
2021 - May 2026 (Anticipated)

Dissertation Title: Modern Advances in Copula Theory: From Generator Construction to Machine Learning Applications in Dependence Modeling

M.S. in Mathematics

University of Louisiana at Lafayette
2019 - 2021

M.S. in Statistics

University of Calcutta, India
2016 - 2018

B.S. in Statistics

Asutosh College, India
2013 - 2016

Minors: Mathematics & Computer Science

Recent Publications

Convergence Guarantees for Gradient Descent in Deep Neural Networks with Non-convex Loss Functions (2025)

Agnideep Aich, Ashit Baran Aich, Bruce Wade
https://doi.org/10.1080/00207160.2025.2522349

Proposes LQCRs, a class of structured regions in deep networks where gradient descent provably converges under standard initialization. This work bridges non-convex optimization and deep learning theory, offering depth-aware guarantees with practical acceleration.

Theoretical Foundations of the Deep Copula Classifier: A Generative Approach to Modeling Dependent Features (2025)

Agnideep Aich, Ashit Baran Aich, Bruce Wade
https://doi.org/10.48550/arXiv.2505.22997

Introduces a generative framework for classification based on deep copula models. This work formulates theoretical guarantees and highlights the role of dependence structures in enhancing prediction performance.

Can Copulas Be Used for Feature Selection? A Machine Learning Study on Diabetes Risk Prediction (2025)

Agnideep Aich, Md. Monzur Murshed, Amanda Mayeaux, Sameera Hewage
https://doi.org/10.48550/arXiv.2505.22554

Explores the integration of copula-based dependency measures into feature selection pipelines. Demonstrates significant improvements in identifying relevant predictors for diabetes risk classification.

Get In Touch

Feel free to send me an email or reach out through any of the platforms above, to discuss potential collaborations and research opportunities.