Mr. Tushar Verma | AI Spectroscopy | Best Review Paper Award

B.Tech student | Indian Institute of Technology Roorkee | India

A motivated undergraduate from the Indian Institute of Technology, Roorkee, specializing in Data Science, Machine Learning and Artificial Intelligence. With hands-on experience building applied ML systems, the candidate has worked on projects ranging from finance-focused temporal modeling to histopathology image analysis and speech emotion recognition. Adept at Python, TensorFlow and PyTorch, they combine strong engineering foundations with research curiosity. Active in campus life, they have held leadership roles in social and technical clubs while supporting junior students as a teaching assistant. Their work emphasizes reproducible experiments, cross-domain generalization, and deploying research-ready prototypes that solve real-world problems.

Publication Profile

Scopus

Education Background

Currently a second-year B.Tech student at the Indian Institute of Technology, Roorkee, the candidate completed higher-secondary education at Heera Lal Public School, New Delhi, and matriculation at St. Mary’s Convent School, Agra. Their coursework and independent study emphasize probability, statistics, linear algebra, signal processing and core programming. They have translated classroom learning into laboratory skills through projects that require time-series modeling, vision-transformer architectures, and audio-signal feature engineering. Regular participation in workshops, hackathons and lab seminars at IIT Roorkee has expanded their practical knowledge and exposed them to contemporary tools such as Transformers, Informer, Spacetimeformer and FinLLama.

Professional experience

Although still an undergraduate, the student has built substantial project experience at IIT Roorkee, collaborating with faculty and peers on multidisciplinary problems. They implemented and benchmarked large language models for financial sentiment extraction, developed Vision Transformer variants for robust histopathology classification and engineered audio pipelines for emotion recognition. Their roles included end-to-end system design: data preprocessing, model training, evaluation, and comparative benchmarking. In campus leadership and coordination positions they organized events, mentored junior students, and contributed to project management and logistics, demonstrating the ability to balance technical responsibilities with team leadership and community engagement.

Awards and Honors

The candidate’s profile highlights project- and community-oriented recognition rather than a long list of formal awards. They have earned praise and institutional support for exemplary project work at IIT Roorkee, gained commendations for leadership roles in student organizations, and received positive assessments as an undergraduate teaching assistant. Their contributions to outreach and mentorship through CRY and campus activities have been noted by peers and faculty. While formal national- or journal-level awards are not listed, the combination of competitive project outcomes, institutional acknowledgements, and leadership responsibilities reflect a trajectory likely to attract future scholarships, internship offers, and academic honors.

Research Focus

Their research centers on applying machine learning to high-impact domains: financial time-series forecasting augmented with LLM-driven sentiment signals, robust cancer prediction from histopathology under staining variability, and audio-based emotion classification. Methodologically they focus on integrating temporal models (Informer, Spacetimeformer) with semantic signals from FinLLama, designing Vision Transformer variants with neighborhood and global attention and KAN classifiers for pathology, and engineering rich spectral and time-domain audio features for robust speech emotion recognition. Emphasis is placed on domain adaptation, cross-domain generalization, reproducibility, and combining deep learning with signal-processing insights for interpretable, reliable systems.

Publication

Combating Antimicrobial Resistance: Spectroscopy Meets Machine Learning

Published Year: 2025

Conclusion

A technically proficient and community-minded undergraduate at IIT Roorkee, they combine solid coursework with ambitious projects bridging ML research and practical applications. Their portfolio demonstrates the ability to design and evaluate novel model architectures, integrate multimodal signals, and deliver reproducible experiments. Leadership and mentoring roles show maturity and collaborative skills, while their project choices indicate a clear trajectory toward research in applied ML for healthcare and finance. With continued publication, internships, and mentorship, they are well positioned to transition from strong undergraduate researcher to competitive graduate applicant or industry research engineer.