The experience.
Native iOS interfaces across commerce, insurance, and fintech. Erafone uses SwiftUI, UIKit, Combine, and SPM; Eraspace uses UIKit, RxSwift, and CocoaPods; JD Sports Indonesia uses SwiftUI and RxSwift.
I build reliable iOS experiences.
Across commerce, fintech, and insurance.
Useful agents.
Grounded decisions.
Human-guided automation.
I’m Engkit, a Senior iOS Engineer exploring applied AI through independent research and hands-on projects. My focus is turning AI capabilities into useful tools for code review, workflow automation, and interaction with agents.
My personal projects include Mobile AI Reviewer, Jira Worklog Automation, and Hermes Agent integration with Telegram. They explore how agents can work with engineering context, prepare reviewable actions, and fit into tools people already use.
I approach AI systems with the same care I bring to mobile engineering: explicit state, testable components, and reliable behavior. My review and automation workflows emphasize source context, structured outputs, validation, fallback strategies, and human approval before posting.
My engineering foundation comes from e-commerce, fintech, and insurance. At PT Erajaya Swasembada, Tbk., I work across Eraspace, Erafone, and JD Sports Indonesia on modular architecture, Swift Concurrency, checkout modernization, and CI/CD.
Previously, I led SalingJaga at Kitabisa and the DANA Promo Squad while assigned through Mitrais. Earlier roles at PT Global Loket Sejahtera and PT Paybill Indonesia round out my experience building and shipping native mobile products.
Connecting AI agents to real engineering workflows through source context, structured outputs, and deliberate control over actions.
Three personal projects exploring AI-assisted review, human-guided automation, and conversational access to agents.
A mobile pull-request review hub with iOS and Android reviewer workflows and Graphify knowledge graphs for architecture and dependency context. Discovery prepares review bundles for reviewer agents. Structured results are checked against the current PR head and diff lines, with duplicate suppression and resumable publication. Telegram notifications accompany the review workflow.
Grounding review findings in source context, validating them against the current diff, and distinguishing source analysis from executed tests.
A Python CLI that turns Google Calendar events into proposed Jira worklogs. Gemini categorizes event titles against configured categories, with a keyword fallback when AI is unavailable. Collection stores proposals for review; posting remains a separate approval step. The workflow includes duplicate checks, payload previews, and multi-project category mapping.
Combining AI categorization with deterministic validation, fallback behavior, and human approval before external writes.
My personal integration project connects Hermes Agent with Telegram, bringing interaction with an AI agent into a messaging interface. It forms part of my independent exploration of how people access and work with AI agents.
Exploring a conversational interface for interacting with an AI agent through Telegram.
Mobile architecture, reliability, and technical leadership across commerce, insurance, and fintech shape how I approach applied AI.
Senior iOS Software Engineer across Eraspace, Erafone, and JD Sports Indonesia. Refactored Eraspace checkout into independent, testable components with explicit state management. Migrated closure-based networking to async/await and actors to improve concurrency safety. Added caching and fallback strategies for product launches and promotions, and revamped the home module with configurable feature toggles and local caching. Designed and deployed on-premises Jenkins CI/CD for iOS and Android. Led Clean Architecture and unit testing initiatives, coordinated Firebase and New Relic monitoring, and introduced knowledge sharing, documentation, and technical assessment practices.
Led the SalingJaga iOS project, Kitabisa’s insurance product, as Tech Lead. Established iOS code style guidelines and contributed to modularization with Swift Package Manager. Helped build the insurance product from the ground up using VIPER, UIKit, unit testing, and CocoaPods. Migrated CI/CD from Bitrise to Codemagic and implemented A/B testing and user segmentation with CleverTap.
Developed DANA iOS features in Swift and Objective-C while working on-site with the DANA team through Mitrais. Led the DANA Promo Squad and contributed to technical assessments and documentation. Built a custom DANA iOS SDK in Objective-C and extended functionality in its internal libraries.
Native iOS interfaces across commerce, insurance, and fintech. Erafone uses SwiftUI, UIKit, Combine, and SPM; Eraspace uses UIKit, RxSwift, and CocoaPods; JD Sports Indonesia uses SwiftUI and RxSwift.
Modular design, explicit state management, and testable components. Swift async/await and actors support concurrency safety; caching, fallbacks, feature flags, and A/B testing support reliability and experimentation.
On-premises CI/CD and mobile monitoring, alongside personal AI research in Telegram agent integration, Gemini-assisted worklog categorization, and code review with source context and validation limits.
Let’s talk about native iOS experiences, modular architecture,
and reliable delivery across commerce, fintech, and insurance.
Regional finalist, Hackathon MDK 2.0 — Emozen app.
Panelist, ITG Career Development Center.