Abhay 
Mundhara
I design and ship AI products that feel immediate: local voice interfaces, offline triage flows, private assistants, and ledgers that map cleanly to how people actually work.
- Private inference
- Voice-first UX
- Operational tools
I build
AI systems voice tools offline apps ledger flows that respond.
- Base
- Belfast
- Mode
- Build, test, ship
- Focus
- AI products
I build software that feels tactile, responsive, and direct. My best work sits where AI, product craft, and operational detail meet: voice OS experiments, offline triage, accounting ledgers, and tools that stay useful when the network disappears.
My work focuses heavily on on-device local AI models, voice interaction UX, and double-entry accounting software designed for actual human workflows, not enterprise checkbox hell.
I started with HTML and CSS, moved through local databases and every flavour of fullstack framework, and kept narrowing in on tools that work without a server, without a cloud, without permission.
Prototype-to-ship workflows for assistants, triage flows, and task-specific AI tools.
React, Node.js, TypeScript, APIs, databases, auth, deployment, and integration work.
Interfaces that stay fast, private, and understandable when latency or network access gets messy.
Ledgers, workflows, dashboards, internal tools, and systems shaped around real user habits.
W O R K
