About
I'm Restu — a technical founder and CTO based in Indonesia. I've spent years building real systems across backend engineering, AI, computer vision, and infrastructure. Now I lead Jenjang, an HR intelligence platform built for the way Indonesian companies actually work.
I don't just architect products. I ship them, operate them, and grow the teams behind them.
Most CTOs come from one deep lane. I've crossed four — backend systems, AI and computer vision, DevOps and infrastructure, and now product leadership. Each phase forced me to understand the layer below it and the layer above. It's why I can talk to engineers, product teams, and investors in the same conversation — and why Jenjang is built the way it is. Kelip.io — a product I led in 2023 that never shipped — is what turned me from an engineer who builds into an engineer who thinks.
Where It Started
My career started where most engineers do: writing backend systems and figuring out how to make them work reliably.
My first real project was a Digital Signature System — embedding cryptographic certificates into PDF metadata so documents could be verified as authentic without changing a single visible character. It was my introduction to building systems where correctness isn't optional.
From there I built Presensiku — an HRIS platform with face recognition attendance that I took all the way to market. Employees check in with a selfie or a wefie; the system verifies their identity and logs their attendance in real time. I launched it during COVID-19, when Indonesian companies urgently needed a way to manage remote attendance and had no time to wait.
Presensiku was my first paid product. The first time real companies depended on something I built. That changes how you think — about uptime, about error handling, about what "done" actually means.
AI at the Edge
Not every customer can afford a data center. That's the problem Edge AI solves — and why I spent 2021 researching how to run computer vision inference directly on device.
Using NVIDIA Jetson Nano and Jetson Xavier NX, I ran real detection workloads at the edge: people counting, vehicle counting, queue detection. No cloud dependency. No bandwidth bottleneck. No infrastructure bill that kills the business case for smaller customers.
This research became the foundation for the full computer vision system I built in 2022 — and the reason it was designed to run close to the camera, not in a distant server room.
Going Deeper
Shipping Presensiku on virtual machines, with Docker and Docker Swarm, was also my first real education in infrastructure. I didn't learn DevOps from a course — I learned it because the product needed it to survive.
That curiosity about building things from the ground up led me into research and development. As an R&D Engineer I built a full Computer Vision System — three YOLO models trained entirely from scratch: object detection, vehicle detection, and license plate recognition.
No pre-trained shortcuts. I collected the data, annotated it, trained each model, evaluated performance, and iterated until it was accurate enough for production in Indonesian road conditions — because public datasets don't reflect local vehicles or plate formats.
The final product gave operators real-time traffic intelligence: vehicle counts by hour and day, most detected plates, type breakdowns — all accessible from a dashboard where users could spin up a new computer vision service without touching a single config file. That project taught me what R&D actually means: you don't just build — you question, test, fail, and rebuild until the thing works in the real world.
First Product Role. Wrong Lesson.
Led Kelip.io — an IoT simulator — as my first real product leadership role. I approached it the way engineers do: identify what to build, build it well. What I missed was the foundation under all of it — understanding what users actually needed, and why.
The product never shipped. Market testing pointed somewhere else. We pivoted to custom IoT projects instead.
That failure reframed how I think about engineering entirely. Technical skill without product thinking is just expensive assumptions. A 10x engineer isn't someone who writes more code — they're someone who understands the problem deeply enough that every line of code actually matters. That's the engineer I decided to become.
Leading Product
Every system I'd shipped had pushed me to care more about how it was deployed, how it stayed up, and how it could scale. That led me into a dedicated DevOps focus — Docker, container orchestration, deployment pipelines, and system architecture designed for resilience.
I stopped thinking about services in isolation and started thinking about systems. How components talk to each other. Where the failure points are. How to build infrastructure that supports the product without becoming a bottleneck.
Now
Everything I built before Jenjang turned out to be preparation for it.
Jenjang is an HR intelligence platform for Indonesian companies — combining an AI-powered ATS, psychology-based candidate assessment, and HRIS in one product. Talent Acquisition teams use it to screen CVs with AI, automate interview emails, run structured psychology assessments, and sort candidates based on real signal — not gut feel.
I also built Jeva — Jenjang's AI virtual assistant, powered by a RAG system and the OpenAI API. Jeva answers HR questions grounded in a company's own data: candidate pipelines, assessment results, employee records. Not a generic chatbot — an assistant that actually knows what's happening inside your organization.
The psychology assessment module is built as a standalone service with its own API — so it can integrate into any external HR system, not just Jenjang. That was a deliberate architectural and product decision: our assessment capability is valuable enough to stand alone.
As CTO, the technical decisions are only part of the job. I run regular 1-on-1 sessions with every team member — not to track status, but to understand them: what motivates them, where they're growing, what's blocking them. Growing engineers who think like builders is as important as any line of code I write.
Shipping beats perfecting.
A product in production, learning from real users, is worth more than a perfect system no one has tested.
Real data beats assumptions.
Whether it's training a model or making a product decision — the answer is always in the data, not the meeting room.
Systems reflect the people who build them.
Good architecture comes from engineers who care. Growing the team is the same work as growing the product.
Build for context, not just requirements.
Indonesian companies have specific needs, workflows, and constraints. Generic software built elsewhere and localized imperfectly is not the same as software built here, for here.
The hardest problems are human, not technical.
Every system I've built eventually taught me this. The engineering is the solvable part. Understanding the people it's built for — that's the real work.
A 10x engineer isn't faster — they're better aimed.
Technical depth alone builds things well. Product thinking builds the right things. The combination is where real impact lives — and Kelip.io taught me the cost of having one without the other.