AI Safety & Alignment
Monitoring and enforcing safe behaviour in models and agents — from refusal circuits to deterministic policy engines.
We pick problems where a measurable system can beat the status quo — then publish the code, the numbers and the failures.
Monitoring and enforcing safe behaviour in models and agents — from refusal circuits to deterministic policy engines.
Auditory interfaces and ASR that close regional accent gaps for Indian-accented speech.
Interpretable machine learning that decodes regulatory signals in non-coding DNA on reproducible, modest compute.
Agent teams that act inside real tooling — with observability and runtime validation built in, not bolted on.
Real-time adaptive optics, acoustic physical layers and fault-tolerant data infrastructure.