Hi, I'm Lukas Kristianto

Lead Mobile Engineer & AI Specialist

Nine years building Android, iOS, and Flutter apps: Tokopedia, sales platforms in three countries, warehouse systems on factory Wi-Fi. Now I lead mobile and AI development at APP Group, where my agent pipelines and document-intelligence services run in production daily.

Contact Me Read My Medium Blog

Professional Profile

Mobile Engineer · AI Engineer · Mentor

I build high-performance apps for Android, iOS, and the web in Kotlin, Flutter, and TypeScript. At APP Group I point machine learning at real operational problems: offline sales across three countries, OCR on delivery notes, tax checks on vendor invoices.

As a mentor I break Android concepts down until the team can ship without me in the room. I modernized legacy codebases and wrote the development standards now used across enterprise systems.

How I lead: through code review and written architecture standards, not meetings. Goals are explicit, plans are shared, and every engineer gets mentoring time.

Best Employee 2024 (APP Sinarmas)

6% Revenue Growth Achieved

40-60% Performance Improvements

My Expertise

Android and Flutter Development9+ years

AI/LLM Integration3+ years

Next.js & Tailwind1.5 years

Team Leadership2 years

Leadership Experience

Lead Mobile Application & AI Development

APP Group (Asia Pulp & Paper)

Jan 2025 - Present

Team Leadership

Lead a team of 4 Android engineers: architecture standards, code review, mentoring, sprint delivery
Define the technical roadmap for mobile + AI initiatives across the group's business units
Partner with Product, UX, and backend teams to define requirements and API contracts
Deliver through the company's project-improvement program end-to-end: registration, go-live, closing
Mentor engineers on Kotlin, Coroutines, and Jetpack Compose performance practices

Technical Achievements

Built an agentic AI system over the company's data lake (SAP, ND6, SQL Server, vector store): leadership reads daily insights and next actions, zero operators
Built an e-invoice & withholding-tax (PPh) verification service: local vision LLM + 20+ rules, ~90 vendors
Built UNSPSC material classification feeding SAP master data, human-in-the-loop with regression-scored changes
Built FX·AI currency-rate capture: 4-tier chain with vision LLM reading bot-walled sources
Launched a Gemini payment-assistant chatbot inside Microsoft Teams
Fine-tuned Gemma to LiteRT for on-device inference; automated internal tracking with Playwright (26 projects)

Senior Mobile Application & AI Development

APP Group (Asia Pulp & Paper)

Apr 2021 - Dec 2024

Key Projects

Sales Force Automation (IDR 10B daily transactions)
ATOWN Internal Social Media (30,000+ APP Sinarmas Employee)
Forestree Plantation System
Tracking Ship/Log Systems

Innovations

Reduced app sizes by 40-60%
Improved performance through Compose
Document Classification with CNN TensorFlow
On-device odometer AI (YOLOv11 + fine-tuned Keras OCR) in VRA Time Sheet, live on Google Play
Standardized Android development practices

Professional Experience

Software Engineer Android

Tokopedia

Apr 2019 - Apr 2021

Performance Optimizations:

Home Page: 5s → 1s load time
Official Store: 4s → 2.3s
Wishlist: 2s load with DiffUtil

Key Achievements:

Reduced APK size by 70% (98MB → 30MB)
Implemented TDD (96% coverage)
Developed Tokopedia Play video feature
Automated Dark Mode for 326 modules

Full-Stack Engineer (Freelance)

Aug 2023 - Present

PT. Rapier Technology

Developed native Android apps with Kotlin and Clean Architecture
Built multiplatform apps with Flutter (BLoC/GetX)
Shipped native iOS apps with Swift and SwiftUI
Created backend services with Golang

Android Engineer

Oct 2022 - Jan 2023

ASTRO

Built chat app with Jetpack Compose
Developed Warehouse Management System
Implemented Android best practices

Flutter Engineer

July 2023 - present

Mutter

Built Flutter Mobile App for Mutter Breast Pump
Maintained and enhanced app features based on user feedback
Stabilized BLE connectivity with pump hardware

Android Engineer

July 2017 - March 2019

RENSEKI

Built POS System app with Java and Kotlin Native
Developed Sales Force Application with Flutter
Implemented Android best practices

Highlighted Projects

Tokopedia E-Commerce

Android

Indonesia's largest e-commerce platform with 100M+ users where I led critical performance optimizations and feature development.

Major Contributions:

Homepage optimization (5s → 1s load time)
APK size reduction from 98MB to 30MB
Implemented Dark Mode for 326 modules
Developed Tokopedia Play video feature
Pioneered TDD (96% test coverage)
Official Store performance improvements

KotlinMVVMGraphQLDynamic Features

40% faster UX on key journeys

Sales Force Automation

Android Native

Multilingual, multi-currency system supporting 3,000+ sales personnel across Indonesia, Malaysia, India, and Thailand with real-time GPS tracking and offline transaction capabilities.

Key Features:

Real-time salesman tracking with geofencing
Offline transaction processing (sync when online)
Multilingual support (4 languages)
Video learning modules for product training
Daily transaction reports with digital signatures

KotlinJetpack ComposeRoom DBGIS

Result: 6% revenue growth | IDR 10B daily transactions

ION Toll (PT Citra Persada Indonesia)

Android + iOS

Toll-road superapp shipped natively on both platforms: a Kotlin multi-module Android app and a SwiftUI iOS app, maintained in parallel by me end-to-end.

Highlights:

24 feature modules on 4 core layers (live traffic, CCTV, rest areas, toll rates)
Payments, wallet & e-money top-ups, digital receipts
SPBU & SPKLU finders, journey hub, news
SwiftUI iOS app with its own unit-test suite

KotlinMulti-ModuleSwiftUIUnit Testing

24 feature modules · 2 native codebases

Screamordance, event companion app

Flutter (client)

Event app for a client built on Rapitech's backend: visitors explore tenants and booths, claim vouchers, and complete missions on Android and iOS.

Engineering details:

Feature-first Clean Architecture with BLoC and get_it DI
httpOnly-cookie sessions with token rotation at the Dio layer
Optimistic favorite UI with rollback and double-tap guards
Indonesian + English localization, Firebase messaging & crashlytics

FlutterBLoCGoRouterFirebase

67 passing tests · 2 languages · Android + iOS

E-Invoice & Withholding-Tax Verification

Agentic AI

Production document-intelligence service for a payment-verification team: vendor e-invoice emails pass 20+ deterministic rules, then a local vision LLM judges document completeness, handwriting, watermarks, and payer match.

Highlights:

Indonesian withholding-tax rules (PPh 15/23/26) code-generated from finance spreadsheets
Charter vs non-charter treatment decided by port of loading
Renders every attached PDF: name consistency, materai, signatures
Deterministic fallback keeps the service running when the LLM is down

PythonFastAPIQwen3-VLpytest

~90 vendors mapped · keeps working without the LLM

Sales-Insight Agent Pipeline

Multi-Agent

An agentic AI system for the group's data lake: it connects SAP, ND6 document archives, SQL Server databases, and a vector store, then turns them into daily insights, recommendations, and next actions that leadership reads every morning.

How it runs:

Five specialist agents draft and cross-check the analysis
A quality reviewer rejects weak output before anyone reads it
Fallback chain from cloud LLMs to self-hosted vLLM (Qwen 14B)
Hybrid BM25 + semantic RAG; caching answers repeat questions at zero model cost

PythonFastAPIvLLMRAG

5 agents · runs daily with no human in the loop

UNSPSC Material Classification

RAG + Rerank

Material-classification system feeding SAP master data governance: a retrieval-and-rerank pipeline assigns UNSPSC codes to procurement items, corrected by human reviewers through an audit queue.

Highlights:

Reviewer corrections baked back into the next pipeline run
Challenge harness regression-scores every rule change
Multilingual evaluation across codebook variants
API server + audit tooling for the review team

PythonRetrievalFastAPIEval Harness

Human-in-the-loop · every change regression-scored

FX·AI, currency-rate capture

Vision LLM

A rate service that never silently fails: central-bank scraping with IDR cross-rate resolution, headful-browser screenshots read by a vision LLM for bot-walled sources, then text-LLM and structured-API fallbacks.

Technical Aspects:

4-tier capture chain, a rate never left uncaptured
Append-only SQLite history (no overwrites)
Scheduled capture via cron, 1-hour cache
Next.js App Router dashboard with per-pair charts

Next.jsTypeScriptSQLitePatchright

Multi-currency · append-only, no silent gaps

VRA Time Sheet — On-Device Vehicle AI

Computer Vision

The group's fleet timesheet app, live on Google Play, reads odometers by camera: a YOLOv11 model detects the gauge, then a fine-tuned Keras OCR model reads digital and analog odometers on cars, trucks, and excavators — all on the device.

What I built:

YOLOv11 object detection for odometer localization
Keras OCR fine-tuned on digital and analog gauges
Works across cars, trucks, and digital-excavator displays
On-device inference: no server round-trip, manual entry cut

YOLOv11KerasTensorFlow LiteAndroid

Live on Google Play

Technical Skills

Mobile Development

Android (Kotlin/Java)
Flutter
Jetpack Components
Material Design
Firebase

AI/ML Development

Multi-agent pipelines
RAG (BM25 + semantic, ChromaDB)
Vision LLM (local Qwen3-VL)
On-device AI (LiteRT, Gemma)
OCR & document intelligence

Web Development

Next.js
Tailwind CSS
TypeScript
Node.js

Other Skills

Team Leadership
Agile Methodology
CI/CD Pipelines
Microservices
Cloud (AWS/GCP)

Get In Touch

Connect With Me

lukaskris12@gmail.com

Phone/WhatsApp

+62 8515-8388562

linkedin.com/in/lukaskristianto

Instagram

@lukaskriiss

Github

lukaskris12

Hi, I'm Lukas Kristianto

Lead Mobile Engineer & AI Specialist

Professional Profile

Mobile Engineer · AI Engineer · Mentor

My Expertise

Leadership Experience

Lead Mobile Application & AI Development

APP Group (Asia Pulp & Paper)

Team Leadership

Technical Achievements

Senior Mobile Application & AI Development

APP Group (Asia Pulp & Paper)

Key Projects

Innovations

Professional Experience

Software Engineer Android

Tokopedia

Performance Optimizations:

Key Achievements:

Full-Stack Engineer (Freelance)

PT. Rapier Technology

Android Engineer

ASTRO

Flutter Engineer

Mutter

Android Engineer

RENSEKI

Highlighted Projects

Tokopedia E-Commerce

Major Contributions:

Sales Force Automation

Key Features:

ION Toll (PT Citra Persada Indonesia)

Highlights:

Screamordance, event companion app

Engineering details:

E-Invoice & Withholding-Tax Verification

Highlights:

Sales-Insight Agent Pipeline

How it runs:

UNSPSC Material Classification

Highlights:

FX·AI, currency-rate capture

Technical Aspects:

VRA Time Sheet — On-Device Vehicle AI

What I built:

Technical Skills

Mobile Development

AI/ML Development

Web Development

Other Skills

Get In Touch

Connect With Me

Send a Message