Local AI systems
Exploring
Running models closer to the machine. Useful when privacy, latency or cost matter more than convenience.
Jayaranjan Jayarajan · Cambridge, UK
Building useful AI.
Building enduring companies.
Founder of Journai and Journai Health. I build products, systems and companies around problems worth solving.
Explore my workCurrently building
Two companies.
One long game.
Journai
AI systems for businesses where generic software isn't enough.
Through Journai I work with teams who need software shaped around their actual constraints: messy data, existing systems, and workflows that refuse to fit a template.
The work spans product building, AI integration, automation and longer technical partnerships. Consulting is part of it. It is not the whole identity.
Explore JournaiSelected focus
- Custom AI inside real business systems
- Product MVPs with technical depth
- Automation that respects existing operations
- Long-term build partnerships
The interesting work lives between the model and the organisation.
Permissions. Edge cases. Data quality. Human review. That is usually where useful AI is won or lost.
Journai Health
Healthcare has plenty of data.
The harder problem is making sense of it.
Journai Health is a more ambitious venture: organising fragmented clinical information into a longitudinal view that care teams can review, verify and act on.
Clinical context
Conditions, encounters, medicines and outstanding actions presented as a coherent patient story, not a pile of records.
Interoperability
Built around FHIR-oriented data structures so information can move between systems without losing meaning.
Explainability
AI-assisted outputs stay linked to source information. Clinicians should be able to inspect and verify.
Responsibility
Clinical oversight, auditability and governance are part of the product direction, not an afterthought.
Selected work
Problems worth building for.
2024Company
CompanyJournai
An AI and software studio building systems for businesses where generic tools fall short.
2025Health technology
Health technologyJournai Health
A clinical intelligence venture focused on longitudinal patient context, interoperability and explainable AI.
2024Selected work
Selected workCustom AI systems
Practical AI tools for professional services and SaaS teams that needed more than a chatbot wrapper.
EarlierFoundation
FoundationBioinformatics & software
A year-long placement working on healthtech and bioinformatics software that shaped later product direction.
Lab
Things I'm exploring.
A quieter side of the work. Prototypes, infrastructure and questions that may become products later, or simply make me a better builder.
ExperimentStatusNote
AI agents in real workflows
Active
Agents are interesting when they can act inside constrained systems. Less interesting as demos.
Hardware & local infrastructure
Ongoing
Software is only half the story. I care about the physical layer: devices, networks, and environments that keep systems honest.
Health technology prototypes
Directed
Small experiments that feed Journai Health. Always with clinical context and governance in mind.
Automation that earns trust
Recurring
The best automation is quiet. It removes friction without inventing new failure modes.
How I think
A few convictions I keep returning to.
01
Stay close to the problem.
Technology is secondary. The work starts with what is actually broken, expensive or unclear.
02
Build before you boast.
A working system beats a polished narrative. Ship something real, then talk about it.
03
Complex underneath. Simple on the surface.
The difficult parts should disappear for the people using the product.
04
Think in years. Ship in weeks.
Long conviction, short cycles. Enduring companies are built that way.
About
I'm Jay.
I'm a founder based in Cambridge. Most of my work sits somewhere between building software, understanding difficult problems, and figuring out whether they can become meaningful companies.
I started Journai to build AI systems for businesses that need more than generic software. Journai Health came from a longer conviction: healthcare technology should help clinicians make sense of information, not add another layer of noise.
Outside the companies, I experiment with local AI, hardware and infrastructure. I like making complicated systems feel simple.
Notes
Writing as the work develops.
Contact
Build something difficult.
If you're working on an interesting problem in AI, software or health technology, I'd like to hear about it.