Four engineers, twenty million dollars a month
A consumer lending platform that underwrites, prices and disburses a loan in about five minutes — and decides seven thousand applications a day without a human reading most of them.
Systems that run whole businesses.Built by specialists.
HummingWave builds payment switches, charging networks, lending platforms and commodity exchanges, and has for thirteen years. Each one is built by an engineering team you could fit around one table.
$900M+Lent through lending platforms we built
19,000+Medical charts coded by AI we built
16,000ATMs on a switch we built
3,000EV chargers under remote management
2013Building since
Some of the companies we build for
YamahaFundMore.aiFinnableLendingkartFYNXTAgSrvIndus Motors GreenIndia1ExicomSmartKarrotNestléMahindraMyntraAkamaiEYHoneywellVerifoneST EngineeringMalayala ManoramaETV BharatGlobal University SystemsPearl AcademyNIRAEmidsMindteck
Practice areas
Six places we go deep
Not a list of everything we could do. These are the areas where we have built the hard parts more than once, and where we would back ourselves against anyone.
01
Lending & credit
Origination, decisioning and servicing. Machine-learned underwriting, alternate-data scoring for thin-file borrowers, and the collections and reporting behind them.
7,000Applications decided daily
02Applied AI
Assistants that act inside a company's own tools, decisions made from incomplete evidence, records read and coded, fleets watched for the faults nobody has named — in production, with a person approving what matters.
5Production systems running AI we built
03SaaS & customer success
Multi-tenant analytics platforms. Tenant isolation, metering, sandboxes, SDKs and the embedded AI that turns account data into something an account manager can act on.
5Continents served by one platform
04E-mobility & energy
Charger remote management, OCPP and OCPI, battery swapping, field engineering. Rule engines that fix faults without dispatching anyone.
2×/minTelemetry per connector
05Payments & banking
Real-time scheme integration, ISO 8583 switching, card management, reconciliation and settlement. Certified against the scheme, not just tested internally.
100,000Transactions an hour at peak
06Marketplace
Exchanges rather than catalogues. Live auctions, price discovery, custody, central clearing and same-day settlement — for commodities, materials and goods.
T+0Settlement, same day as trade
How this works
What four engineers built
Every firm claims depth. Here is what ours actually produced.
4
engineers built the whole lending platform
- $20M disbursed every month
- 7,000 applications auto-decisioned daily
- 1.7% delayed payments
- Five-minute loans, running as a service for two lenders
Not four people plus an offshore team. Four engineers, with test and business analysis alongside them.
A small team that goes deep costs less in total, because it spends its time building rather than coordinating, and the person who designed your system is the person writing it.
We would rather stay small and deep than large and broad. That is a real constraint on what we take on, and it is the reason our engagements run for years rather than months.
Work
Case studies with the hard parts left in
Architecture, constraints, the alternatives we rejected and why. Clients are described rather than named, with their agreement — except where they have asked to be named.
—
—
Two frameworks, one dashboard
A customer success platform competing with Salesforce, serving tens of thousands of users across five continents — and the architectural decisions that let a small team keep pace with a company a thousand times its size.
—
The cheapest repair is the one nobody drives to
A remote management platform for three thousand EV chargers worldwide — where the design goal is not to detect faults faster, but to resolve them without dispatching anyone.
—
Sixteen thousand ATMs, most of them where the roads run out
Our client runs about seven per cent of India's ATM estate, largely in tier-4 and tier-5 towns. We build and run the switch integration, the UPI and IMPS stack, and the reconciliation underneath.
Engineering notes
Written by the people who built it
One technical point each, made completely. No ghostwriters, and nothing selling at the bottom.
12
An LLM for medical coding that says what is missing
Made to choose a code the chart could not support, the model in the medical coding engine we built picked the most common one. So the engine flags the missing detail instead of guessing, and a coder asks the clinician. It has coded more than 19,000 anaesthesia charts.
11
Put the LLM last: automating a finance review queue
In the finance review queue we built, the source system, rules and past reviewer decisions answer each transaction first; an LLM handles only what they cannot. It sees only the fields it needs, returns only a client and a reason, and leaves the line for a person when it cannot answer.
10
Three ways to isolate a tenant in a NoSQL SaaS platform
For a multi-tenant SaaS platform on a document store, we weighed an account per tenant, a table per tenant and shared tables, and chose shared tables. Sharding tenants across databases and enforcing tenant scope in the data layer bought the isolation back.
Sectors
And the places we have been before
Ninety-odd clients across thirteen years. If your industry is here, we have already learned its vocabulary — which is usually the first month of any engagement.
AgricultureAutomotiveCapital marketsEducationHealthcareLogistics & commoditiesMediaRetail & e-commerceEnterprise & consulting
Talk to us
Tell us what is hard about it
We are most useful on systems where something is genuinely difficult — scale, regulation, integration, latency, or a decision nobody has been able to automate. If that is the conversation, it will be with engineers, not an account manager.