Capabilities
We take on project-based work in identity, private communication, and AI systems — usually where those three meet and the off-the-shelf answer does not hold. Scope is agreed up front, and your team owns the result.
Self-sovereign identity
W3C DIDs and verifiable credentials, AnonCreds, holder wallets, issuance and verification services, and the trust registry decisions that surround them.
- DID methods & resolution
- AnonCreds & selective disclosure
- Wallet SDKs, mobile and web
DIDComm messaging
Encrypted, transport-agnostic messaging between agents and organisations, including mediators, routing, and the operational work of running them.
- Agent & mediator deployment
- Protocol design
- Interop testing
AI agents
Agent pipelines built on Ojas: retrieval over private data, tool use with real limits, observability, and evaluation before anything reaches users.
- Retrieval & indexing
- Evals & regression suites
- Tracing & cost control
Privacy engineering
Threat modelling, key management, and architectures where the operator structurally cannot read the data — then the review that proves it.
- Threat models
- Key custody & recovery
- Architecture review
How an engagement runs
Scope
Threat model, constraints, and what success has to look like.
Prototype
A working slice of the hardest part, not a document about it.
Build
Shipped in increments, with evals and traces from the start.
Transfer
Your team runs it. Documentation and key custody included.
Engagements are project-based, with scope agreed before we start. We take a limited number per quarter.
Start a conversation