Benshi.ai SDK - AI/ML Data Ingestion & Intervention Platform

[SDK]

Tech Stack

Kotlin, React Native, Javascript, Typescript, Java

Organization(s)

Benshi.ai

Resources

Product Overview

Benshi.ai SDK - AI/ML Data Ingestion & Intervention Platform

Project Details

The Challenge

Benshi.ai, a non-profit backed by the Bill & Melinda Gates Foundation, needed a robust, lightweight SDK to bridge partner applications with its AI/ML platform. The goal: capture meaningful behavioral data from real-world app usage and deliver personalized, adaptive interventions in resource-constrained regions, without adding friction or overhead to partner development teams.

The Solution

As lead developer, I architected and built Benshi.ai's cross-platform SDK from the ground up a modular data ingestion and intervention engine powering the feedback loop between partner apps and Benshi.ai's reinforcement learning platform. Deployed across partner applications spanning social media, e-commerce, e-learning, and healthcare access in underserved markets, the SDK was the invisible backbone enabling AI-driven personalization at scale.

What I Built

Behavioral Data Ingestion: Captures app lifecycle events, screen-level usage durations, and domain-specific user interactions across diverse partner verticals including healthcare, e-commerce, e-learning, and social media

Intervention Engine: Delivers context-aware nudges via push notifications, and in-app messages, with auto-logging user responses (engaged, dismissed, or blocked) to continuously feed the AI/ML feedback loop

Android SDK: Built in Kotlin with WorkManager for efficient, battery-aware background task scheduling and event ingestion. Published on Maven Central

React Native SDK: TypeScript wrapper bridging the native Android SDK for React Native partner apps, centralizing data and network management while accelerating integration timelines. Published on NPM

Modular Architecture: Segmented into purpose-built modules: Core, E-Commerce, E-Learning, Loyalty, Payments, Community Health, Social Media; keeping package sizes minimal and integrations clean across diverse partner use cases

Partner-First Documentation: Delivered comprehensive, use-case-specific integration guides tailored to each partner's domain, reducing onboarding friction and time-to-integration

Impact

Deployed across global partner platforms in underserved regions, the SDK enabled Benshi.ai's AI/ML engine to deliver personalized healthcare interventions at scale, directly supporting the Bill & Melinda Gates Foundation's mission to improve health outcomes in low-resource environments.