Explore
Featured work
We build
to last.
We build
for the Goals.
We build
in sixty days.
00 What we do
Anything scores well on a laptop. We build the part that comes after — the thing still working in a clinic with intermittent power, on a floor of fifteen-year-old machines, in an office that accounts for every rupee.
Seven sectors.
One studio.
Idea to working MVP
0days
Not a prototype. Something real users are in, with the instrumentation to prove it moved.
What we bring
Built inside the LEAF incubation space at GLS University, Ahmedabad, and grant-backed by the university.
Start a brief01 Sixty days
Day 01 — 07
We go to where the problem is. Not a workshop — the ward, the shop floor, the field. Then we map what actually happens against the target you are accountable for.
What comes out
Day 08 — 21
We design the system before we write it: what the data model has to carry, where agents hand off to people, and which decisions must still make sense to an auditor in two years.
What comes out
Day 22 — 48
Web, mobile, models, and hardware in fortnightly increments you can open and use. Nothing gets demoed from a slide.
What comes out
Day 49 — 60
Live with real users, measured from the first day, with a scale roadmap that names the policy and procurement gates you will actually hit.
What comes out
02 Capabilities
Everything we build is the same chain — something senses, something reasons, someone decides, something is written down. What changes between jobs is how much of that chain you need.
01 60 days
Discovery on site, an architecture that survives audit, and a pilot real people are using at the end of it — the whole chain, stood up in sixty days.
SenseReasonDecideRecord
02 Assistants
Assistants that adapt to how a person or a small team already works, instead of asking them to reorganise their week around a tool.
SenseReasonDecideRecord
03 Virtual COO
For places where a wrong call costs something measured in people. Every escalation runs through a human who can answer for it, and everything lands in the ledger.
SenseReasonDecideRecord
04 Platforms
Sensing through to dashboard, built so the number in the quarterly report can be traced back to the device that produced it.
SenseReasonDecideRecord
03 Sectors
A clinic with intermittent power is not a staging environment. These are the sectors we know well enough to design for their bad days, not their demo days.
Good health and well-being
Early-warning systems, triage support, and analytics for clinics, hospitals, and field NGOs working where the nearest specialist is three hours away.
What we build
Quality education
Adaptive learning, assessment tooling, and skill mapping for institutions that need to prove progress rather than attendance.
What we build
Decent work, industry, responsible production
Industrial IoT and predictive maintenance for small and mid-sized producers, retrofitted onto the machines already bolted to the floor.
What we build
Sustainable cities and communities
Smart building systems, planning tools, and infrastructure monitoring for cities that have to do more with the grid and the water they already have.
What we build
Decent work and reduced inequalities
Risk assessment and inclusion platforms for lenders serving people with no credit file but plenty of evidence they are good for it.
What we build
Peace, justice and strong institutions
E-governance, transparency tooling, and citizen services built to survive an audit and a change of administration.
What we build
Zero hunger, responsible production, climate action
Yield forecasting, traceability, and sensor networks for cooperatives and producers who lose margin between the field and the mandi.
What we build
04 Next
A portable six-in-one device and the software behind it, for catching malnutrition and growth risk in children early. It replaces the tape, the scale and the MUAC band with one instrument, and turns a screening into data at the point of care.
From the screening to the caseload
Screening Height, weight and mid-upper-arm circumference in one pass
At the point of care The reading lands as a record, not a note to type up later
Against the standard Plotted where the WHO growth standards put it
Caseload Every child screened, and who needs seeing
05 Work
The same plate, at half the size. Two are live and running; two are not finished, and are shown as what they are rather than dressed up with a picture of a product that does not exist yet.
An AI and IoT app for detecting child malnutrition, built for the Anganwadi workers who do the screening and the parents who take the child home. Registration, growth screening, supplement stock and the vaccination schedule, in one place.
Read the project
Heart attack prevention and response in one app: a personalised risk assessment that asks the questions a clinician would, real-time monitoring, preventive guidance and a route to emergency help.
Read the project
A quiz platform for schools and colleges. Teachers build, publish and mark assessments; the results come back as something readable — by class, by topic, by student — rather than a column of scores.
Visit the site
An open dataset platform for India: find public data, publish your own, and have the schema and metadata checked before either happens. Lineage is kept from source to curated to published, so a number in a report can be traced back to the file it came from.
Visit the site06 Custom builds
Most of what we ship is ordinary commercial work — dashboards, operations systems, internal tools, sites. Same team, same sixty days, same standard. Here is some of it.
Ordering, dispatch and stock for a rotomoulding plant, on one screen instead of four spreadsheets.
Billing, purchase returns, inventory and customer history for an eyewear chain, in one system.
A patient lifecycle and pharmaceutical record a clinician can actually work through between appointments.
Daily sales, 1,457 medicine categories and expiry tracking, for a counter that cannot stop to reconcile.
Portfolio value, cashflow and brokerage across 215 managed properties, for brokers in Ahmedabad.
Hardware operations wired to an ERP and a set of custom automations that run without being watched.
A few numbers that matter
Trusted by teams like yours
07 Studio
The person who sat in the clinic is the person who writes the model. Nothing gets thrown over a wall, because there is no wall to throw it over.
All four, one room LEAF Incubation, GLS University Ahmedabad, India
01
IoT integrator · Data analyst
Designs the layer where connected devices stop producing noise and start producing decisions worth acting on.
Builds Device networks, telemetry pipelines
LinkedIn
02
AI/ML engineer · Full stack
Takes a model from a notebook to something that survives real traffic, real edge cases, and a network that keeps dropping.
Builds Domain models, agent orchestration
Portfolio
03
UI/UX designer · Frontend
Makes the interface legible to someone using it for the first time, under pressure, on a small screen, in a hurry.
Builds Interfaces, design systems
Portfolio
04
Research · Content strategy
Does the reading before anyone writes code, so what we build answers the question that was actually asked.
Builds Field research, sector briefs
LinkedIn08 Questions
Short answers, in plain words. If yours is not here, send it with your brief and we will answer it properly.
We are an AI and IoT innovation studio in Ahmedabad, India. We build systems that sense, reason, decide and keep a record — for clinics, classrooms, factories and public offices — and we also take on ordinary commercial work: websites, dashboards, CRMs and mobile apps.
Sixty days, from a problem worth solving to something in use: seven days of ground truth on site, fourteen days of architecture, twenty-seven days of build in fortnightly increments, and twelve days of pilot with real users.
In the LEAF Incubation Co-Working Space at GLS University, Ellisbridge, Ahmedabad, Gujarat 380006. The studio is grant-backed by the university, and we work with clients across India and internationally.
Seven: healthcare, education, manufacturing and MSMEs, urban systems, finance and inclusion, government, and agriculture and rural. Those map to UN Sustainable Development Goals 2, 3, 4, 8, 9, 10, 11, 12, 13 and 16.
No. Most of what we ship is ordinary commercial work — websites and product sites, dashboards and internal tools, CRM and operations software, web and mobile apps, AI automation and agents, and MVP builds. Same team, same sixty days, same standard. You do not need a mission statement to hire us.
Our AI and IoT platform for child health and nutrition. A portable six-in-one device replaces the measuring tape, the weighing scale and the MUAC band, and the software reads growth and nutrition risk against WHO growth standards — so early signs of malnutrition are caught at the screening rather than at the next visit. Built for parents, doctors and frontline health workers. It lives at sanrakshya.in.
Send a brief with the problem, the constraint and the deadline, plus any documents you already have. A person reads it — usually the same day — and comes back with an approach, what we would need, and what we think it costs. If it is not a fit we will say so.
Four people, all in one room: Vansh Balyan on IoT and data, Ashwin Solanki on AI/ML and full stack, Dhruv Solanki on UI/UX and frontend, and Aastha Sangtani on research and content strategy. The person who sat in the clinic is the person who writes the model.
09 Start
Send the problem, the constraint, and the deadline. Skip the pitch — we would rather read the awkward specifics.
A person, not a router. Usually the same day.
How we would tackle it, what we would need, and what we think it costs.
No sequence, no drip. If it is not a fit we will say so.
Got it. Someone will actually read this.
One of the four of us will read it and come back to you — usually within a working day. If it is urgent, mail us directly.
