Aditya Singla
From zero research experience to Best Oral Presentation at an international IEEE conference in Tokyo
Where Aditya Started
His Background
- • Junior at Cupertino High School in Silicon Valley
- • Strong interest in AI and machine learning
- • Curious about cloud computing and system reliability
- • No prior research or publication experience
His Goals
- • Conduct original AI research with real-world applications
- • Publish in a peer-reviewed venue
- • Present at an international conference
- • Build expertise in neural networks and anomaly detection
The Problem He Wanted to Solve
"Cloud systems power everything from banking apps to social media. When they fail, millions of users are affected. I wanted to build AI that could detect these failures before they cause real damage— catching issues that current methods miss."
— Aditya, before joining YRI
The Research
Working with his YRI mentor Vamika Perumal from IIT Madras, Aditya developed a novel approach to anomaly detection in large-scale cloud infrastructure. His research addressed a critical challenge: existing methods either miss real anomalies or generate too many false alarms.
A Novel, Multistep Neural Network Approach to Improve Anomaly Detection in Online Operational Status
Cloud system failures affect millions of users; current detection misses critical issues
IBM Cloud Console telemetry - 39,365 data points with 117,000+ features
Two-step neural network pipeline with progressive filtering
Detected 16/25 anomalies vs. 6/25 in prior research (2.7× improvement)
Novel Two-Step Pipeline Architecture
Unlike single-model approaches, Aditya's system uses two specialized neural networks working in sequence. The first network identifies suspicious patterns across the massive feature space, while the second examines flagged instances more closely to separate true anomalies from false alarms.
Algorithm to identify the most predictive variables
Flexible threshold-based filtering between stages
Modular framework adaptable to different methods
The Outcome
Best Oral Presentation Award
International Conference on Advances in AI and Machine Learning
Tokyo, Japan • March 20-22, 2026
IEEE Conference Proceedings (peer-reviewed)
Best Oral Presentation in his session
Aditya presenting his research at IEEE AAIML 2026 in Tokyo, Japan
Presenting at an international IEEE conference in Tokyo was surreal. Winning Best Oral Presentation for my session made the trip even more special. I would like to thank my research mentor, Vamika, for her guidance throughout this journey.
High school junior interested in AI, no research experience, no publications
IEEE published, Best Oral Presentation at international conference in Tokyo
Why This Research Matters
Improvement in anomaly detection rate compared to prior research
Features analyzed from real IBM Cloud infrastructure data
Framework applicable to cloud systems serving millions of users worldwide
Ready to Start Your Research Journey?
Join the YRI Fellowship and work with expert mentors to conduct original research, publish in top venues, and present at international conferences.
Apply Now