Manini Pay
Tekkon Technologies Pvt. Ltd.
0 to 1
Product built from ground up
300-500%
Market growth during COVID-19 period
Multi-bank
Integration with card processors and banking APIs
ML-powered
Fraud detection and merchant analytics
Overview
Manini Pay was a cloud-native digital wallet and payments platform I helped build from the ground up at Tekkon Technologies, an Australian-Nepali product services company. The platform served both merchants and consumers, defining core money movement flows including wallet top-ups, merchant payments, payouts, and refunds, with integrations to card processors and banking APIs.
This project launched during a pivotal moment for digital payments in Nepal. The COVID-19 pandemic drove 300 to 500 percent growth in digital payment volumes across the country, making 2020-2021 the most transformative period for fintech in Nepal's history. Competing against established players like eSewa (with 10 million or more users) and Khalti required strong product differentiation and technical excellence.
The platform was designed as a microservices architecture from day one, reflecting the cloud-native approach that would later become the standard for fintech products in the region. This was not just a consumer app — it was a complete payments infrastructure including event schemas for ML pipelines, traceable audit logs for regulators, and developer-friendly APIs for merchant integrations.
The Problem
Nepal's banking API landscape was fragmented and non-standardized. Each bank had its own API format, authentication mechanism, and settlement process. Building a payments platform required integrating with multiple banks individually while maintaining a consistent user experience. Additionally, the market was competitive with established players who had significant head starts, requiring both technical innovation and smart go-to-market strategy to gain traction.
My Role
Product Manager
I led product discovery and launch of Manini Pay, defining core money movement flows and data collection schemas. I partnered with engineers, data scientists, and account managers to pilot predictive models for merchant churn and high-value merchant identification. I also owned dashboards and KPIs for payments and ML experiments, using insights to refine training data, risk thresholds, routing logic, and checkout flows.
The Approach
We adopted a bank adapter pattern to abstract away the complexity of individual bank integrations. Each bank's unique API was wrapped in a standardized interface, allowing the core payment orchestration logic to remain bank-agnostic. This meant adding new bank partners required writing only the adapter layer, not modifying core systems.
I defined data collection and event schemas so that transaction, KYC, and behavioral data could feed ML pipelines for fraud detection, risk scoring, customer lifetime value estimation, and merchant segmentation, while maintaining traceable audit logs for partners and regulators.
Developer experience was a priority. I created developer-friendly documentation, Postman collections, and sample apps in React and Python that showcased how to integrate core payment APIs, handle webhooks for settlement and refund events, and consume AI-enhanced endpoints.
Key Features
What we built
Cloud-Native Payments Engine
Microservices-based payment orchestration supporting wallet top-ups, merchant payments, payouts, refunds, and settlements.
ML-Powered Fraud Detection
Machine learning models for real-time transaction scoring, anomaly detection, and risk-based transaction routing.
Merchant Analytics
Predictive models for merchant churn and high-value merchant identification, informing onboarding prioritization and support.
Developer API Toolkit
Comprehensive REST APIs with Postman collections, sample integrations in React and Python, and webhook support for real-time event notifications.
Real-Time Dashboards
Payment status and risk scores exposed through APIs and webhooks, enabling internal tools and merchant dashboards to react in near real time.
Multi-Bank Integration Layer
Bank adapter architecture abstracting the complexity of Nepal's fragmented banking API landscape into a consistent interface.
Tech Stack
Key Lessons
What I took away from this project
Building from zero to one requires ruthless prioritization — every feature must justify its existence
Developer experience is a competitive moat for payment platforms
In fragmented banking landscapes, the abstraction layer is the most valuable infrastructure
Audit-ready data architecture should be built from day one, not retrofitted
More Projects
Explore other work
Want to build something similar?
I help companies scale their products and build high-performing teams. Let's discuss how I can help with your next project.
Get in Touch