FinTech · Embedded lending
Merchant Lending Platform
A platform connecting SMBs with lenders, matching every loan request to the best available offer automatically.
The challenge
Loan processing was only partially automated. Merchants applied, and matching them to a suitable lender product relied on manual work.
The platform introduced loan offerings: lenders publish products with an amount, interest rate and duration, and each incoming request is matched to the best fit. It also had to manage tenure-based variable rates and per-merchant limits.
Our Founder's role
- Built the merchant-facing React application from scratch, including its client-side state-management design, within an eight-person team.
- Implemented seamless sign-in. The app is embedded in a parent product, so merchants already logged in there never see a second login.
How it works
- 01
Merchant frontend
A React application embedded in a parent product, authenticated by token, talking to both the lending core and the centralised KYC platform.
- 02
Lending core
A Ruby on Rails service owns the business rules (loan offerings, rates and merchant limits) in PostgreSQL.
- 03
Offer matching
Each new loan request is assigned the best lender offering based on facility limit, merchant confidence score, turnover, business type and other signals.
- 04
Verification and review
Identity checks run through the centralised KYC platform and its third-party provider. Reviewers give final approval in the lending core.
- 05
Notifications
Sidekiq jobs on Redis email merchants about approval status and alert admins to new loan requests and completed KYC approvals.
Engineering decisions
Separate what changes fast from what must stay stable
- Core lending logic stays on Rails and PostgreSQL. Verification lives in a dedicated, shared KYC service. Each scales and evolves independently.
- Identity verification is delegated to a specialist provider rather than rebuilt in-house.
No second login
- Token-based sign-in carried over from the parent product removes a drop-off point at the very start of the funnel.
Outcomes
- Loan requests matched to lender offerings automatically instead of by hand.
- Merchants move from parent product to loan application without re-authenticating.
- Lending rules and verification scale independently.
Technology
- React
- Merchant frontend
- Ruby on Rails
- Lending core
- Go
- Centralised KYC service
- PostgreSQL
- Loan offerings, settings and limits
- Kafka
- KYC status events
- Sidekiq · Redis
- Background jobs and notifications
- Tilaka
- ID proofing, OCR, liveness
Contact
Have something you want to build?
Tell us what you're building, or what you'd like to automate. Let's figure out the next step.