Jandas Corporation · AI-Native Engineering · Los Angeles
Custom AI agents. Engineering automation. Production software.
Jandas helps companies integrate AI into the way their engineering organizations actually work. We design and build custom agents for software development, QA, DevOps, production operations, knowledge systems and business workflows — connected securely to your code, cloud infrastructure, APIs, data and engineering tools.
AI Agents · LangGraph · LangChain · RAG · MCP
20+ years of production engineering experience. Now applied to AI-native software organizations.
From agents that improve how your team works to complete AI-powered products, we architect the intelligence layer and every system it needs to perform in production.
AI that does work — not another chatbot.
We design production AI agents that reason across your company’s tools, data and workflows, then take authorized action with human oversight where it matters.
Give your engineers AI coworkers.
Connect agents to GitHub, Jira, CI/CD, documentation, cloud, monitoring and team communication — automating repetitive work across development, review, testing, release and operations.
Production AI requires much more than an LLM API.
We design the architecture underneath AI-native systems: APIs, compute, data, identity, queues, events, observability, deployment and security.
Build AI into the product itself.
We architect and build complete AI-powered products across mobile, web, backend and cloud — from a focused feature to an entire new platform.
Production depth behind every AI architecture.
More than two decades across consumer, enterprise and connected systems means our AI is built by engineers who understand the software and hardware it must operate inside.
Agents can coordinate work across the systems your team already uses, without adding another destination to monitor.
Your tools don’t need another dashboard. They need an intelligence layer between them.
AI Engineering Workflow Assessment
We analyze development workflows, source repositories, CI/CD, QA, DevOps, cloud, documentation, engineering tools and production operations — then identify where AI can create leverage without introducing unmanaged risk.
Assess my engineering workflowFind the high-value workflows where AI automation can remove friction and create measurable impact.
Design agents, models, tools, data access, human controls and security boundaries as one system.
Connect agents to the code, cloud, data and engineering tools your organization already relies on.
Measure accuracy, reliability, cost, safety and business impact before expanding autonomy.
Monitor, improve and expand agents safely as models, tools and organizational needs change.
Agents are designed around how your engineers already build, test, deploy and operate — not around an isolated AI demo.
Identity, permissions, data access, evaluation, observability and failure modes are part of the system from day one.
We define clear approval boundaries so agents can move work forward without taking unmanaged action.
Accuracy, reliability, cost and engineering impact stay visible as agents learn, expand and move into production.
Two decades of shipping and operating real systems for companies like Meta, Walmart, Disney, GE and Etsy — and our own AI-run product, LineFlip. That production depth is what our agents and automation are built on.
Human-networking AI platform
Recent
Currently shaping an AI product that helps people build real professional relationships — combining social-system design, mobile experience and applied AI into one coherent platform.
Meta — Oculus / Horizon
Entertainment VR
Ran the Oculus Move initiative end to end — new features, critical fixes, roadmap and eventual sunsetting — while keeping the FitnessTracker app alive on Quest. Audited Horizon apps for security and reverse-engineered performance bottlenecks in Horizon TV and Meta Horizon mobile.
LineFlip — LineFlip.net
Now live on iOS
Designed, built and launched end to end under Jandas Corporation — the iOS app now live on the App Store, alongside the publishing tools, payment gateway and cloud infrastructure. The engagement-operations platform was architected from day one so AI-assisted workflows can handle the repetitive work as volume grows: listings publish, demand is analyzed, guests are engaged and followed up, and payments reconcile — with people approving the decisions that matter. It is the same agent-and-automation architecture we build for clients, running on a live product.
MidWestTape — hoopla
Technical Assessment
Ran a comprehensive technical assessment across hoopla's iOS and Android apps, engineering workflows, QA processes and production operations. Audited source code, architecture, release pipelines, security posture and team workflows, then sized AI feature opportunities and delivered a prioritized roadmap with concrete remediation steps.
Walmart — Sam's Club
Enterprise retail
Untangled performance hotspots in the Sam's Club app and led the iOS migration to Walmart's Glass platform, wiring GraphQL into the stack so the client could finally move at store scale.
Magnopus — Expo 2020 Dubai
Entertainment AR/VR
Architected the bridge system powering the Expo 2020 Dubai iOS and Android app, then built AR prototypes, camera composers and 3D visualization tools around it. Added a Parasol CI/CD pipeline with testing, linting and automated deploys on every pull request.
GE Power
Enterprise
Built the data, networking and serializer foundations of the APM iOS app from the ground up, then added middleware that made Field Vision faster in the field and integrated a thermal camera SDK for industrial inspections.
AXS / Veritix
Ticketing
Rebuilt the core architecture of the Access ticketing system, shipped AXS app v2 and the Red Rocks venue app, and designed a Proximity SDK that triggered geo-aware notifications at the right moment.
Etsy
E-commerce
Made Etsy's Seller and Buyer apps work for more people — implementing accessibility across both iOS and Android so the marketplace stayed usable and inclusive at scale.
Disney Interactive
Entertainment
Built Disney Mobile Connector Tech and wired it into games across the portfolio, tuned Unity titles for iOS and Android, and built the CMS and content-pipeline systems that kept those games fresh.
Napster (Rhapsody)
Music / Streaming
Crafted the custom UI controls, offline caching layer and core audio engines behind Napster on iPhone and iPad, then led the Android phone, tablet and Google TV clients.
Omnicircuit
Connected hardware / Fleet
Built an entire fleet-tracking product from zero: device firmware, Linux TCP/UDP servers, web dashboard, PostGIS geocoding backend and a covert vehicle camera module — hardware, cloud and interface in one loop.
Architecture decisions, agent workflow maps, evaluation results, model records, security boundaries and runbooks are developed with the system so teams can operate it with confidence.
System diagrams, data models, API contracts, event flows and decision records that explain why the system is shaped the way it is.
Auto-generated and human-edited reference docs, SDK guides and example flows that reduce integration time for client teams.
Incident response, deployment, rollback, observability and on-call procedures tested against the real system before go-live.
Feature specs, user flows, PRDs and user-facing guides that keep product, engineering and support aligned.
Prompt libraries, model cards, evaluation results and agent workflow maps that make AI behavior auditable and repeatable.
Threat models, access-control matrices, audit trails and compliance mappings prepared alongside implementation.
Jandas Corporation is an AI-native engineering and software architecture firm led by Andrew Choi, a software architect with more than 20 years of experience building production systems.
We help companies transform traditional engineering workflows into AI-native ones by designing custom agents, engineering automation and cloud architectures that work with the tools and systems teams already use.
Unlike AI consultancies built around demos and prompts, Jandas brings decades of production experience across mobile, backend, cloud, connected systems, CI/CD, enterprise architecture and large-scale consumer products for organizations including Meta, Disney, Walmart, Etsy, GE, AXS, Napster and hoopla.
Languages
Swift, Kotlin, Java, C/C++, Objective-C, Embedded C, TypeScript, JavaScript, PHP, Python, SQL
Frameworks & SDKs
iOS, Android, AOSP, React, React Native, Flutter, Unity, Node, Firebase, AWS, GCP, Azure
AI
LangGraph, LangChain, MCP, RAG, tool calling, multi-agent workflows, evaluation
Build & DevOps
Jenkins, Fastlane, GitLab, Azure DevOps, CircleCI
Databases
MySQL, PostgreSQL (PostGIS), Firebase Realtime / Firestore
Tell us how your team currently builds, tests, deploys and operates software. We’ll identify where custom agents can remove repetitive work, accelerate delivery and give your engineers more time to engineer.