Back to all work

Case study 02 / Inova Platforms / 2023 - 2025

Built a production AI platform across 7 models, 4 data APIs, and 38 cloud functions.

James architected and operated a full-stack AI product across a SwiftUI client, Python microservices, GCP, Firebase, subscriptions, observability, and automated distribution workflows.

01

Context

DeepChamp AI is a full-stack sports analytics product spanning a native consumer application, AI research and reasoning services, production infrastructure, subscriptions, notifications, and analytics.

02

Mandate

Own the product and technical system end to end, from the customer experience and AI pipeline to cloud delivery, monetization, reliability, and growth operations.

03

Product and platform

Built a SwiftUI iOS app and Python microservices. Architected a three-stage AI pipeline spanning Perplexity, GLM-4.5, and Gemini, with seven AI models and four sports-data APIs overall.

04

Production operations

Deployed 38 cloud functions across GCP and Firebase. Added CI/CD, Crashlytics, Cloud Logging, Firebase services, subscriptions through Superwall and StoreKit 2, and automated distribution workflows.

05

Result

Turned a multi-model AI concept into a live, instrumented production platform with a native customer surface, cloud services, monetization, and operating controls.

06

Leadership lesson

AI implementation is a systems discipline. Model quality matters, but so do orchestration, data freshness, reliability, product state, monetization, observability, and ownership.