|
Manjil Nepal | मन्जिल नेपाल I'm a final year computer science student at SRM University-AP, India. |
|
|
Research Interest: Federated Learning, Multimodal Deep Learning, Computer Vision |
Experience |
|
Research Intern - DREAM:Lab, Indian Institute of Science, Banglore
Feb 2025 - Present
Advisor:
Prof. Yogesh Simmhan
Research Area: Federated Learning and Edge Accelerators |
|
Research Intern - Agency for Science, Technology and Research, Singapore
Aug 2025 - Jan 2026 Research Area: Federated Learning (Finance), Generative Models (Medical Imaging) |
|
Summer Research Intern - Indian Institute of Technology- Dhanbad, India
May 2025 - July 2025
Advisor:
Prof. Tamoghna Ojha
Research Area: Federated Learning for Resource Constrained IoT Devices |
|
Peer Reviewer - Scienctific Reports
June 2025 - Aug 2025 |
ResearchI'm interested in federated learning, computer vision, deep learning and generative AI. Some papers are highlighted. |
|
DPxFin: Adaptive Differential Privacy for Anti-Money Laundering Detection
via Reputation-Weighted Federated Learning
Renuga Kanagavelu, Manjil Nepal, Ning Peiyan, Cai Kangning, Xu Jiming, Fei Gao, Yong Liu, Goh Siow Mong Rick and Qingsong Wei ACM International Conference on AI in Finance (ICAIF)- Workshop, 2025 (Oral Presentation) paper / code Developed DPxFin, a federated learning framework for Anti-Money Laundering that uses reputation-guided adaptive differential privacy to dynamically adjust client noise based on model alignment, protecting sensitive financial tabular data while preserving accuracy and achieving a stronger privacy–utility trade-off with improved robustness to data leakage under IID and non-IID settings. |
|
Modeling Student Sentiment from Academic Feedback: A Machine Learning
Approach
Manjil Nepal, Sunit Soni, M Krishna Siva Prasad, Jayash Shrestha, Prashant Dhimal THE 16th IEEE ICCCNT, 2025 (Oral Presentation) paper / code Developed an NLP-based sentiment analysis system to automatically classify large volumes of student feedback as positive, neutral, or negative for educational quality assessment. Applied text preprocessing and count vectorization, and trained Naive Bayes, KNN, and. Enabled institutions to extract actionable insights from unstructured feedback, supporting data-driven improvements in instruction and curriculum. |
Achivements |
|
3rd Place in Financial Survival Benchmarking
Manjil Nepal, Parth Kiran Hanchate, Fei Gao, Qingsong Wei, Renuga Kanagavelu FinSurvival Challenge, ACM-ICAIF, 2025 paper / code The challenge centers on time-to-event prediction, modeling how long it takes from an initial event (e.g., loan issuance) to an outcome (such as repayment or liquidation). Models are evaluated using the Concordance Index (C-index), where 0.5 indicates random performance and 1.0 represents perfect ranking. We secured 3rd place with a C-index of 0.8472. |
Miscellanea |
|
Member & Researcher - Next Tech Lab, AP, India
Oct 2024 - Present |
|
Senior Member - AIESEC at Amaravati
Mar 2025 - Aug 2025 |
|
Source code from code link |