# Kyle C.

> Building AI tools that make engineers faster | iOS & Android | LLM Agents, MCP

Location: San Francisco, California, United States
Profile: https://flows.cv/kylec

I build tools that make engineers faster.

I built an agentic crash triage system saving 100+ minutes per person per week, an LLM pipeline for test coverage analysis, and a recommendation engine using semantic embeddings. Lately deep in MCP and Claude Code.
Mobile background in iOS and Android. Shipped three apps to the App Store using SwiftUI, CoreML, and on-device ML.
Open to Developer Productivity, AI Tooling, and senior iOS roles.

## Work Experience
### Software Engineer II: Infrastructure @ Scribd, Inc.
Jan 2025 – Present
* Built an AI crash-triage agent integrating Firebase, Datadog, Jira, and Slack across 4 production mobile apps (iOS + Android), eliminating 100+ minutes/person/week of manual work.
* Architected a locally hosted open-source LLM pipeline: custom scripts extract iOS XCTest code into structured JSON, the LLM analyzes coverage gaps, and auto-generates dynamic test coverage reports.
* Prototyped an XCUITest recommendation engine using semantic embeddings of Swift test code to surface context-aware suggestions based on similarity, an early RAG application for developer tooling.
* Integrated Google MLKit OCR into Android Espresso to enable vision-based UI automation on dynamic screen content.
* Authored MCP-based skill definitions and agentic workflows using Claude Code to accelerate internal developer tooling and automation pipelines.

### Senior SDET @ Scribd, Inc.
Jan 2022 – Jan 2025 | San Francisco Bay Area
* Ported Apple Vision ML library into XCUITest to power image-recognition-driven test automation.
* Trained a YOLOv3 CoreML model tailored to app-specific UI elements, enabling robust image-based regression testing on iOS.
* Expanded iOS XCUITest framework and release automation coverage using modern Swift concurrency (Async/Await), reducing manual regression effort.
* Developed a locally hosted LLM service with custom tooling to extract, analyze, and document test coverage, improving visibility across the QA org.

### Self Employed @ Self-Employed
Jan 2020 – Jan 2022
Released three iOS apps in App Store:

* Phigment: an experimental ML creativity app for image processing and AI-generated text, built with SwiftUI, MVVM, and TensorFlow Lite for fully on-device inference.
* Rijkys: a museum discovery app for browsing Rijksmuseum's digital art collection, built with SwiftUI, Combine, and MVVM with advanced search filter and RESTful JSON integration.
* Skylob: a multimedia explorer for NASA's vast digital archive of images, videos, and YouTube content, built with SwiftUI, UIKit, and Combine.

### Senior Mobile SDET @ Fitbit (now part of Google)
Jan 2017 – Jan 2020 | San Francisco, CA
* Built and maintained iOS (Swift/XCUITest) and Android (Java/Kotlin Espresso) test frameworks for smart wearable device testing in lab and Firebase environments.
* Developed CI tooling and reporting services to support continuous integration pipelines for mobile releases.

### Software Engineer In Test II @ if(we)
Jan 2014 – Jan 2016
* Led Android automation team; scaled test framework across Appium, Gradle, and TestNG with Jenkins CI pipeline.
* Built cross-platform automation infrastructure, Jira API tooling, and Selenium Grid web automation in a page-object model.
* Winner, Engineering Innovation Hackathon Q2 2016.


## Education
### Bachelor of Science in Computer Science
University of California, Riverside


## Contact & Social
- LinkedIn: https://linkedin.com/in/kylechihcheng

---
Source: https://flows.cv/kylec
JSON Resume: https://flows.cv/kylec/resume.json
Last updated: 2026-04-05