Featured buildBuatLapak — AI storefront builder
ShippedI build AI-native apps, agents, and automations for the work your team keeps repeating
Next.jsTypeScriptReactTailwind CSSGSAP motionLLM appsAI agentsRAG systemsWorkflow automationGenerative mediaAPI integrationREST APIFastAPITanStackAI avatarAI voiceNode.jsPythonCI/CDGitVercelCloudflareProduct designNext.jsTypeScriptReactTailwind CSSGSAP motionLLM appsAI agentsRAG systemsWorkflow automationGenerative mediaAPI integrationREST APIFastAPITanStackAI avatarAI voiceNode.jsPythonCI/CDGitVercelCloudflareProduct design
A few things I have been building.
Public product surfaces and staging builds, described by the problem, the product path, and the current status. Open a card for the fuller build context.
View project A public product for Indonesian UMKM to turn a business name into a professional web page, product showcase, and WhatsApp-ready customer path without coding.
- Problem
- Many local businesses still depend on marketplace links or WhatsApp Status and do not have a professional page they control.
- What I built
- A no-code flow around business discovery, generated copy and layout, Google Maps context, product content, SEO, and WhatsApp checkout.
- Stack
- Next.jsGoogle MapsAI generationWhatsApp
- Status
- Shipped.
- Why it matters
- It gives local businesses an owned web presence without asking them to learn web design or build a full commerce stack first.
View project A staging product for AI-led livestreams that answer chat, react to gifts, and support automated selling on TikTok and YouTube without the operator appearing on camera.
- Problem
- Live selling normally asks a host to stay on camera, answer viewers, and keep the session moving in real time.
- What I built
- A conversion-focused product surface around AI streamer personas, live loops, natural AI voice, photorealistic avatars, multi-platform streaming, and flexible plans.
- Stack
- Next.jsAI avatarAI voiceTikTok LiveYouTube Live
- Status
- In Development.
- Why it matters
- It makes always-on live commerce possible without requiring a human host to stay in front of the camera for every session.
View project A hiring workspace for growing HR teams to publish roles, review candidate evidence, run structured interviews, approve decisions, and keep every candidate informed.
- Problem
- Hiring context gets split across spreadsheets, interview notes, chat threads, and follow-up reminders, making the next step hard to own.
- What I built
- A connected workflow for role requirements, candidate records, evidence review, structured scorecards, human approvals, and candidate updates.
- Stack
- Role workflowCandidate evidenceScorecardsHuman approvalsCandidate updates
- Status
- Shipped.
- Why it matters
- It keeps the hiring story together so teams can make clearer decisions without rebuilding context from spreadsheets, chat threads, and inboxes.
How I can help your business.
Bring a messy workflow, rough AI idea, or content pipeline. I turn it into a clear product direction, useful automation, or launch-ready MVP with the business context kept in view.
Clear scopeReviewable AI outputFast prototypeProduction handoff
Turn manual work into AI software
Document intake, review flows, reporting, CRM updates, admin panels, and internal tools that replace repeated copy-paste work.
Shape raw ideas into working MVPs
Product scope, interface flow, prompt logic, validation states, database shape, and deployable Next.js builds.
Build content and media pipelines
Image, video, voice, avatar, ad creative, batch asset generation, and approval workflows for repeatable production.
Connect AI to your business stack
OpenAI, Gemini, Anthropic, REST APIs, FastAPI, Node.js, Python, Supabase, Stripe, email, webhooks, Vercel, and Cloudflare.
From heavy-equipment marketing to building useful AI products.
MPRBA previously worked as a freelance marketer in the heavy-equipment industry, where research, positioning, operational context, and practical communication mattered every day.
In March 2026, that business-grounded background became the base for AI product building: finding real problems, shaping lean workflows, and shipping tools people can actually use.
From brief to product, in three steps.
A tight, transparent process. You always know which stage we're in and what's coming next.
Scope
Clarify the problem, users, data, and the smallest useful version worth proving.
Problem + source mapBuild
Prototype the flow, connect the model, and test the edge cases before launch.
Tested build in stagingLaunch
Ship a usable product, document the handoff, and keep improving from evidence.
Live surface + next iterationWhat I can build with AI.
From generative media to operational software, the focus stays on systems that make a real workflow clearer or faster.
AI web apps
SaaS MVPs, dashboards, portals, admin panels, and focused internal tools.
Generative media
Image, video, voice, avatar, ad creative, and repeatable content pipelines.
AI agents
Research, support, sales, ops, and knowledge-work agents with reviewable outputs.
Workflow automation
Document intake, reporting, CRM updates, notifications, and exception handling.
RAG and API systems
Source-grounded search, OpenAI, Gemini, Anthropic, REST APIs, and FastAPI.
Product design and launch
User flows, prompt UX, validation, testing, deployment, and handoff.
Have a real problem, a rough idea, or an internal workflow that needs a better tool?