Rabby Ilyas

(About)

Hi, I'm Rabby.

Muhammad Rabby Ilyas. A full-stack & AI engineer from Indonesia , building AI products for founders abroad and a startup of my own at home.

Rabby's avatar, waving
Indonesia · GMT+7Binus · Computer ScienceFounder @ JuraganinClients in Remote since 2024Still learning: data science

(Go on, throw the stickers)

(Story)

I started in data, automating influencer and SEO reports. Then machine learning. Then the whole product around it. Today I work directly with founders, and I'm building one of my own.

I am a full-stack software engineer focused on AI-driven SaaS products. Most of my recent work uses Next.js, Supabase, Pinecone, Gemini, FastAPI, OCR, and NLP to turn messy workflows into clean product experiences.

I work directly with founders and client teams, which means I care about the product shape as much as the implementation: what needs to be automated, what should stay simple, and how the system can stay maintainable after the first version ships.

(How I work)

Three rules I don't break.

1

Practical first

AI earns its place by removing a real step from someone's day. If a plain form does the job, it gets a plain form.

2

Evidence over vibes

Evals in CI, gates that refuse bad saves, scores that trace back to a source. I like being able to prove it works.

3

Own the whole thing

Schema, prompt, UI, infrastructure, the 3 a.m. alarm. I care what happens after the demo.

(Education)

Computer Science

Bina Nusantara University

Algorithms, data structures, databases, machine learning: the foundations everything on this site is built on.

(Experience)

Where I've been shipping.

  1. Jul 2026 — now

    Founder

    Juraganin

    An AI store builder for Indonesian small businesses. Two agents, one runs the shop, one writes the site.

  2. 2025 — now

    Freelance full-stack & AI engineer

    Clients in the US & Germany

    Second Act Pro, Daphne and AtlasPad: matching, reasoning spaces and realtime voice.

  3. Aug — Nov 2025

    Full-Stack Developer

    Growthmarketing.tw

    A bilingual SEO AI Overview dashboard; 2,000+ keywords, reports about 4x faster.

  4. Mar — Aug 2025

    AI/ML Engineer

    Comms8

    ML experiments, influencer data enrichment, full-stack tooling and QA with teams in London, Hong Kong and Tokyo.

  5. May 2024 — Feb 2025

    Data Intern

    Comms8

    Automated reporting and data pipelines for influencer, SEO and campaign analytics.

(Quick answers)

The short version.

Who is Muhammad Rabby Ilyas?
A full-stack and AI engineer from Indonesia. He builds AI products for founders in the United States and Germany, including Second Act Pro, Daphne and AtlasPad, and is building his own startup, Juraganin.
What does Rabby build?
Production AI products: tool-calling agents with approvals and evals, realtime voice agents over WebRTC, retrieval and matching systems with scores a human can explain, and the full-stack SaaS around them, from the database to deployment.
What is his tech stack?
TypeScript, Next.js and React on the front end; Python, FastAPI, Supabase and PostgreSQL on the back end; Pinecone, OpenAI (including the Realtime API), Gemini and OpenRouter for AI; three.js for 3D; AWS (Fargate, Pulumi) and Vercel for infrastructure.
Where is he based? Does he work remotely?
He is based in Indonesia (GMT+7) and works remotely with teams abroad: currently founders in the United States and Germany, and earlier teams in London, Hong Kong and Tokyo.
Is Rabby available for new projects?
He is open to one more serious build alongside his current clients and Juraganin. The fastest way to reach him is email at muhammadrabby7@gmail.com or booking a call at cal.com/mrabbyilyas/call.
Where did he study?
Computer Science at Bina Nusantara University (Binus) in Indonesia.

(Currently learning)

Still a student of it.

Associate Data Scientist in Python

In progress

Continuing the DataCamp track to strengthen statistics, modeling, and Python data workflows.

  • Python
  • Pandas
  • Statistics
  • Machine Learning

Production RAG and Vector Search

Active practice

Learning by building client systems with Pinecone, legal/document retrieval, embeddings, and grounded Q&A.

  • Pinecone
  • Embeddings
  • Retrieval
  • Evaluation

AI Audio and Session Intelligence

Client-driven learning

Exploring live audio, speaker mapping, transcript pipelines, and AI summaries for long sessions.

  • Deepgram
  • Realtime APIs
  • Transcripts
  • Summaries

Better SaaS Architecture

Ongoing

Refining how I structure Next.js, Supabase, background jobs, dashboards, and deployment boundaries.

  • Next.js
  • Supabase
  • Async jobs
  • Vercel