Federated Learning✦Health AI✦Privacy-Preserving ML✦Financial Inclusion✦DevOps AI✦Edge Inference✦Low-Resource AI✦Biomedical AI✦Legal Access AI✦Decentralized Compute✦Community Health✦Economic Inclusion✦Federated Learning✦Health AI✦Privacy-Preserving ML✦Financial Inclusion✦DevOps AI✦Edge Inference✦Low-Resource AI✦Biomedical AI✦Legal Access AI✦Decentralized Compute✦Community Health✦Economic Inclusion✦

AI Research for Underserved Populations · Est. 2026

AI systems that

Federate.

Nepal-first AI research lab building privacy-preserving, federated, and applied AI systems for underserved populations — and the world.

View research →Get in touch

Research Tracks

Founding Members

Est. Kathmandu

Independent

About

Global methods.
Proven
in Nepal.

Aiventra is an independent research laboratory building privacy-preserving, federated AI systems for low-resource, high-stakes environments — proven in Nepal and South Asia, designed to scale globally.

We operate at the intersection of fundamental research and real-world deployment — building AI infrastructure that works for hospitals, banks, governments, and communities that can't afford to share their data to benefit from AI.

Privacy-PreservingFederated AIApplied ResearchIndependent

How we sustain

Services & Consulting

AI strategy consulting, contract research, and government digital transformation projects across Nepal and South Asia.

Education & Training

Corporate AI workshops, annual intensive schools, online certifications, and university curriculum partnerships.

Products & APIs

Licensed APIs for South Asian NLP and health AI models, SaaS tools, curated datasets, and white-label AI systems.

Startup & Venture

Incubating AI startups from research, spinning out commercial ventures, and licensing novel AI methods globally.

Lab Milestones

A phased path to research independence

Year 1Foundation

Consulting · Corporate training · Community events · Memberships

Year 2Stabilization

API licensing · Online courses · Data products · Conferences

Year 3Scale

SaaS products · Startup incubation · Government contracts

Year 4+Independence

Spinouts · Patent royalties · White-label tools · Endowment

Research

Four
Research Tracks

Each fundable independently.
All interconnected by design.

01Core track

Decentralized & Privacy-Preserving AI

Federated, sovereign, and trustworthy AI for institutions that can't share raw data.

Federated LearningDifferential PrivacyDecentralized ComputeDigital Identity & Trust

Why this matters

Hospitals, banks, and government bodies hold sensitive data they cannot legally share. Federated AI lets them collaborate on models without exposing a single record — the foundational technology layer every other track runs on.

Core methods

Federated learning · Differential privacy · Secure multi-party computation · On-device inference · Verifiable AI outputs

02Core track

Biomedical & Human Health AI

AI for health in contexts where data is scarce, sensitive, and culturally specific.

Biotech / Drug DiscoveryNeuroscience & Mental HealthClinical AICommunity Health AI

Why this matters

The highest-grant-potential track with the strongest human impact story. All sub-domains share a core challenge: making AI work on scarce, sensitive, culturally specific medical data — a direct consumer of the privacy infrastructure in Track 1.

Core methods

Medical imaging AI · NLP for clinical notes · EEG/BCI signal decoding · Multimodal biomedical models · Drug-herb interaction prediction

03Applied track

AI for Economic Inclusion

Making formal systems — banks, courts, schools — accessible to those locked out of them.

Financial InclusionEducation AILegal Access AIInfrastructure Assessment

Why this matters

The widest funder base of all four tracks. The unifying logic: AI making formal systems accessible to people currently excluded from them. Shared methods across every domain — NLP, alternative data modeling, and voice-first interfaces for low-literacy users.

Core methods

NLP for local languages · Alternative-data credit models · Voice-first conversational AI · Document intelligence · Computer vision for infrastructure assessment

04Revenue track

DevOps & AI Infrastructure

Applied AI infrastructure for emerging market tech companies — the track that funds the other three.

AI Incident ResponseCI/CD IntelligenceAI CybersecurityEdge / On-device AI

Why this matters

Hundreds of IT and SaaS companies across South Asia need DevOps AI but are priced out of enterprise tools. A region-focused SRE AI platform generates consulting and SaaS revenue that cross-subsidizes the research tracks — targeting cash-flow positivity by Year 2.

Core methods

Agentic AI · Anomaly detection · Log analysis & root cause analysis · Model quantization · Edge deployment optimization

How the tracks connect

Four tracks. One coherent lab.

Track 01Foundational technology layer

Provides the privacy infrastructure that Tracks 2 and 3 run on. Every model trained without sharing raw data starts here.

Track 02Flagship research track

Highest grant potential, strongest human impact story, deepest Nepal anchor. The face of the lab externally.

Track 03Breadth track

Widest funder base, most visible to government and development organizations. Keeps the lab's mandate broad and relevant.

Track 04Revenue engine

Generates income from South Asia's IT sector to buy research independence from pure grant dependency.

Team

Research-led.
Mission-driven.

Aiventra is built by practitioners — researchers and engineers who run experiments, write the code, and deploy AI in the environments that need it most.

Core team

Prabhash Kumar Jha

Co-Founder & Executive Director

Researcher in medical imaging, federated learning, and NLP for low-resource languages. Building open-source AI tools for underserved populations in South Asia.

LinkedIn ↗GitHub ↗HF ↗

Rodan Maharjan

Co-Founder

Co-founded Aiventra with a shared vision of building an independent, Nepal-first AI research laboratory solving real problems through applied research.

LinkedIn ↗

Debesh Jha

Lab Director & Senior Research Advisor

Professor and research leader guiding the lab's scientific direction. Brings deep expertise in AI research and international academic collaboration.

Scholar ↗ORCID ↗GitHub ↗

We're building the team.

Open to researchers, engineers, and collaborators working on federated AI, health AI, and applied systems for South Asia.

Agentic AI · Anomaly detection · Log analysis & root cause analysis · Model quantization · Edge deployment optimization

How the tracks connect

Four tracks. One coherent lab.

Track 01Foundational technology layer

Provides the privacy infrastructure that Tracks 2 and 3 run on. Every model trained without sharing raw data starts here.

Track 02Flagship research track

Highest grant potential, strongest human impact story, deepest Nepal anchor. The face of the lab externally.

Track 03Breadth track

Widest funder base, most visible to government and development organizations. Keeps the lab's mandate broad and relevant.

Track 04Revenue engine

Generates income from South Asia's IT sector to buy research independence from pure grant dependency.

AI systems that