Home Research Domains

Research Domains

Where Theory Meets the Himalayas.

Contextual

Offline-First AI

def optimize_constraints():
  # Silicon Valley builds for Cloud.
  # We build for the Edge.
  
  connectivity = "Intermittent"
  
  if internet_fails():
      run_inference_locally()
      
  return "Healthcare for All"

We research Edge Computing to turn a nurse's $100 smartphone into a world-class radiologist. Our AI works where the internet stops, bridging the last mile of care.

Linguistic

Sovereignty

def train_national_llm():
  # Preserving our identity
  langs = ["Nepali", "Newari", "Tamang"]
  
  for lang in langs:
      # Don't translate. Contextualize.
      embed_cultural_nuance(lang)
      
  return "Voice of the Nation"

Global LLMs are deaf to us. Our NLP research creates datasets and architectures for our 120+ languages, ensuring our culture survives and thrives in the digital age.

Framework

Governance

import buddhist_principles as dharma

def audit_algorithm(model):
  # Ethics is a technical constraint.
  
  if model.is_black_box():
      raise TrustError("Explain!")
      
  dharma.apply_ahimsa(model)
  return "Wise AI"

We define a new standard for AI Governance by fusing global technical standards with Eastern philosophy. We prioritize privacy (Federated Learning) and trust (XAI).

We develop efficient language and vision models and apply scientific machine learning to physics, ecology, and health.

Our work blends theory, computation, and open science.

Scientific ML Model Efficiency Language Processing Computational Physics Optimization Autonomous Systems

Research Areas

Our research spans multiple domains of AI and ML, from scientific discovery to real-world optimization.

Scientific Machine Learning

Bridging Physics and Data

"Do LLMs have the ability to solve novel research problems in SciML?"

SciML PINN NeuralODE UDE Symbolic Regression

Language Processing

Giving Voice to the Voiceless

"Does repetitive translation create loss in meaning while using LLMs?"

LLM Fine-Tuning Reasoning Bias & Misinformation XAI

Physics AI

Modeling the Physical World

"Can we predict droplet impact dynamics and fluid flows using SciML?"

PINN NeuralODE UDE

Chemistry AI

Accelerating Discovery

"Is it possible to predict effective chemical reaction pathways using ML?"

Symbolic Regression SciML UDE

Model Efficiency

Doing More with Less

"Can we build a useful SLM with a very small dataset?"

SLM Attention Heads MOE Inference

Healthcare AI

Frugal Diagnostics

"Can we build a comprehensive epidemiological model for COVID using ML?"

XAI Inference

Optimization

Solving Complex Logistics

"Can we generate timetable schedules without conflicts using multi-constraint optimization?"

Reinforcement Learning MOE

Autonomous Systems

Navigating the Himalayan Gap

"Can we build autonomous agents that navigate GPS-denied, high-altitude terrain?"

SLAM Sensor Fusion Reinforcement Learning Computer Vision

Collaborate with Us

Have a novel research problem? Let's work together to solve it.

Propose a Project