Building ML systems & fintech infrastructure · Palo Alto + remote

Technical founder & builder. I ship ML systems, computer-vision hardware, and fintech infrastructure for emerging markets.

Marcha, Stanford ’29, Computer Science and Management Science & Engineering. I design and ship production systems end to end: a deployed computer-vision waste sorter (JUNKIES, HSBC Climate Innovation Award 2024), a credit-decisioning fintech for unbanked Indonesian SMEs (TrackEco), and AI-native recommendation experiments at Runway Avenue. Olympiad medalist (Gold, ICYS Economics 2022).

Building Runway Avenue · TrackEco Studying Stanford ’29 · CS + MS&E Stack Python · C++ · CV · LLMs

Selected work marchasr@stanford.edu

2 ventures founded · technical 0→1
1 deployed CV system · JUNKIES
32+ SMEs underwritten · TrackEco
10+ press features across Indonesia

/ thesis

What I’m building toward.

2026 →

The next generation of category-defining products will come from founders who can move fluently between deeply technical execution and the on-the-ground reality of underserved markets.

01
Fintech for the unbanked majority

Sixty million Indonesian SMEs are invisible to formal credit. The opportunity isn’t another neobank; it’s a new credit signal. TrackEco is the first expression of that idea.
02
AI-native commerce

Recommendation, search, and onboarding are being rebuilt on LLMs and embeddings. The teams that design these surfaces from first principles will define the next decade of consumer software.
03
ML in the physical world, not just the cloud

The next decade of high-impact ML is at the edge — on-device inference, sensors, and physical systems deployed where the work actually happens. JUNKIES is the first instance: a computer-vision classifier shipping at the trash can, not the cloud.

/ selected work

Things I’ve started or led.

Two ventures founded · 2022 → present

Fintech Infrastructure · Data Systems

TrackEco

Role: Co-founder · Head of Research & Eng
System: Waste-stream telemetry → credit-decisioning engine
Stack: Python · Postgres · Bank-rail APIs · React Native
Market: 67.8M tons / year · Indonesia waste

A fintech platform for the 60 million Indonesian SMEs the formal credit system can’t see. I designed the data model that turns waste-stream telemetry into a fundable credit signal a bank can underwrite, plus the integration layer that pipes those signals into partner-bank rails. Shipped to 32 SME pilots across East Java, with a roadmap to 200+ in the next year.

“Community-based waste management needs to be expanded. TrackEco is scaling our network from restaurants and malls to schools — to manage far larger volumes.” — Marcha, in iNews Surabaya / TIMES Indonesia

Press · 4 outlets

ANTARA News iNews Surabaya TIMES Indonesia Tribun Surabaya

Indonesia's Minister of Tourism Sandiaga Uno demonstrating JUNKIES — KOMIB's AI-powered automatic waste sorter — by feeding a plastic bottle into the unit.

Demo · Menparekraf Sandiaga Uno

Computer Vision · Edge ML · Hardware

KOMIB & JUNKIES

Role: President · Tech Entrepreneurship Div.
Built: Smart Trashcan 2.0 · multi-class CV classifier
Pipeline: Data collection → training → on-device inference
Recognition: HSBC Climate Innovation Awards 2024

Led KOMIB’s Tech Entrepreneurship Division and built JUNKIES — a deployed computer-vision waste sorter (Smart Trashcan 2.0) classifying plastic, paper, metal, and organic waste in real time. The full pipeline ships end-to-end: dataset collection, model training, real-time inference on edge hardware, and live operations in pilot facilities. Demoed internationally to Indonesia’s Minister of Tourism & Creative Economy, who endorsed roll-out plans for airports, Borobudur, and Labuan Bajo.

/ experience

Internships & roles.

2023 — present

Jan 2026 Present
Product Development Intern
Runway Avenue · Fashion-Tech
- Own the experimentation loop end to end, from user research and PRDs through shipping and measurement, partnering with engineering on AI recommendation features.
- Instrument AARRR funnels and run cohort retention analyses to surface friction and prioritize experiments against a north star metric.
- Research LLM native recommendation systems and agentic social commerce primitives to inform roadmap bets.
May 2025 Aug 2025
Financial Advisory Intern
KPMG International
- Modeled cross border tax structures for multinational clients, pressure testing optimizations against shifting regulatory regimes.
- Synthesized tax policy research into client ready recommendations, owning the analysis end to end alongside senior advisors.
- Shipped compliance playbooks that translated regulatory constraints into actions the client’s ops team could execute against.
Apr 2024 Jul 2024
Operations Intern
World Wildlife Fund
- Designed a renewable energy adoption strategy for rural communities near Bukit Barisan National Park, owning the work from baseline data collection through deployment plan.
- Authored sustainable tourism proposals to protect biodiversity hotspots including Komodo National Park, scoped against on the ground partner capacity and shipped to field teams.
Dec 2023 Jan 2024
Finance Intern
Bank Central Asia
- Built profitability models across retail banking portfolio, surfacing growth levers in savings, lending, and investments.
- Shipped reusable reporting templates that cut analysis time from hours to minutes and became the team’s default workflow.

/ honors & awards

Recognition.

Six awards · 2021–2024

2024

HSBC Climate Innovation Acceleration Awards

Climate innovation · High School & University track

Winner
2024

Astranauts Award

National startup & innovation competition · Indonesia

2nd Runner-Up
2023

Young Economist Competition — Royal Economic Society

International economics research

Honorable Mention
2022

International Conference of Young Scientists

Economic division · International research conference

Gold Medal
2021

Indonesia Economics Olympiad

National olympiad · Economics

National Finalist
2021

ASEAN Young Entrepreneurs Award

Regional · Southeast Asia

Awardee

/ education & toolkit

Where I study. What I build with.

Stanford University

Class of 2029

B.S. Computer Science and Management Science & Engineering · GPA 3.88 / 4.00

Coursework: applied machine learning, optimization, stochastic models, statistical inference, data structures, linear algebra. Stanford BASES · Women in Investing · Product Pathways.

Languages & systems

Python
C++
SQL
Git
Postgres
React Native
REST APIs

ML / AI toolkit

Computer vision
Edge inference
LLMs & embeddings
Retrieval / RAG
Agent design
Eval design
Prompt engineering
Claude Code
Cursor

Research & product

A/B testing
Cohort analysis
Statistical inference
Optimization
Financial modeling
PRDs & technical specs

Languages

Indonesian — Native
English — Fluent
Mandarin — Business

/ get in touch

Let’s connect.

Open to conversations on product, fintech, AI-native commerce, and emerging markets — or anything thoughtful that’s worth building. Email is the fastest way to reach me.

Technical founder & builder. I ship ML systems, computer-vision hardware, and fintech infrastructure for emerging markets.

Fintech for the unbanked majority

AI-native commerce

ML in the physical world, not just the cloud

TrackEco

KOMIB & JUNKIES

Product Development Intern

Financial Advisory Intern

Operations Intern

Finance Intern

HSBC Climate Innovation Acceleration Awards

Astranauts Award

Young Economist Competition — Royal Economic Society

International Conference of Young Scientists