CURRENT PROJECTS
Active Research
Explore our current research initiatives making real-world impact
AI Research Facility
Active
2025-2026
Smart Blind Curve Alert: Preventing Accidents
IoT Lab
Nepal's mid-hills face frequent road accidents at blind curves due to narrow roads, steep slopes, and poor visibility, with existing measures like signs proving insufficient. Our research proposes a "Smart Blind Curve Alert" system that detects vehicles approaching from both sides and provides real-time audiovisual warnings to drivers. This solution aims to prevent collisions, reduce deaths and injuries, and minimize economic hardship for families. This project is supported by the Grant from MoEST, Nepal.
Active
2024-2026
Multimodal Audio LLMs: State-of-the-art audio-driven language models integrating speech recognition and synthesis with contextual AI
Covering audio-language integration, speech recognition and synthesis, multimodal AI architectures, and contextual audio understanding systems.
Active
2024-2026
Structure-Based Virtual Screening and Activity Prediction of Millions of Molecules Using Machine Learning
AI /DS Research Facility
Covering computational drug discovery, virtual screening at scale, molecular activity prediction, and machine learning for pharmaceutical research.
Active
2024-2026
Nepali Large Language Model: Customized NLP model fine-tuned on Nepali datasets for localized AI applications and digital inclusion
AI /DS Research Facility
Covering Nepali language modeling, low-resource NLP, cultural adaptation of AI, and digital inclusion through localized language technology.
Active
2024-2026
Healthcare Multimodal LLMs: Integrating clinical text, imaging, and speech data with deep learning for precision diagnostics and patient care
AI /DS Research Facility
Covering multimodal healthcare AI, clinical data integration, precision diagnostics, and deep learning for comprehensive patient care.
Active
2025-2027
Efficient LLM Serving through Hierarchical Speculative Decoding with Adaptive Routing
AI /DS Research Facility
Covering large language model optimization, speculative decoding techniques, adaptive routing algorithms, and efficient model serving architectures.
Active
2025-2027
AI-Assisted Retopology for High-Fidelity 3D Face Reconstruction
AI /DS Research Facility
Covering AI-assisted mesh optimization, high-fidelity 3D modeling, facial reconstruction techniques, and automated retopology algorithms.
Active
2025-2027
Continuous Sign Language Recognition (CSLR) on NSL Dataset
AI /DS Research facility
Covering continuous sign language recognition, Nepali Sign Language dataset processing, gesture recognition, and real-time translation systems.
Active
2025-2027
Enhancing Drug-Induced Liver Injury Prediction: Using Multi-Modal Graph Neural Networks with Explainable AI
AI /DS Research Facility
Covering drug safety prediction, multi-modal graph neural networks, explainable AI techniques, and pharmaceutical toxicity assessment.
Active
2025-2027
Deep Learning Framework for Flood Prediction and Flood-Induced Food Security Risk Assessment in Koshi Province, Nepal
AI /DS Research facility
Covering flood forecasting models, deep learning for disaster management, food security risk assessment, and climate resilience planning.
Active
2025-2027
Importance-Driven Temporal Memory Architectures for Large Language Models: Prioritizing Important Past Events Over Recent Noise
AI /DS Research Facility
Covering temporal memory mechanisms, attention optimization, long-context processing, and relevance-based prioritization in language models.