const engineer =
Muhammad Ramadiansyah
AI & Data Solution Provider
Co-Founder & CTO at Intura · Jakarta, Indonesia
I build AI products end to end — from data pipelines and ML systems to the interfaces people actually use.
the stack I build with
// about
Engineer across the full AI stack
From data pipelines and ML models to the product surface — I like owning problems end to end.
I'm an AI & Data Solution Provider and current Co-Founder & CTO at Intura, with 8+ years across machine learning engineering, data science, and full-stack development.
I've shipped production AI at scale — recommendation engines, real-time data platforms, forecasting systems, computer vision, and LLM tooling — moving fluidly from Kafka pipelines and cloud infra (AWS & GCP) to React front-ends.
I care about systems that hold up in production: observable, cost-efficient, and measured by real business impact like transaction uplift, churn reduction, and infrastructure savings.
- Based inJakarta, Indonesia
- CurrentlyCo-Founder & CTO at Intura
- FocusRecSys · LLMs · Data Platforms
- M.Sc.Computer Science, Univ. of Indonesia
- Years in Data & AI Engineering
- 8+
- AI/ML apps in production
- 30+
- As an entrepreneur
- 2 Years
- Students taught in data & software engineering
- 1000+
Years in Data & AI Engineering
AI/ML apps in production
As an entrepreneur
Students taught in data & software engineering
// experience
Where I've shipped
Eight years building and operating production AI — from e-commerce recommendation to LLM platforms.
Feb 2025 – PresentCo-Founder & CTO
Intura Pte. Ltd. · Singapore
Building an AI Visibility & Brand Research platform and the data/LLM infrastructure behind it, serving B2B clients across fashion, retail, SaaS, edutech, and beauty.
- Built a Brand Research & AI Visibility platform that evaluates how brands surface on AI platforms by generating and analyzing user queries.
- Built a Social Media Listening platform with dashboards and scraper pipelines for TikTok, Instagram, Threads, and X (Twitter).
- Built a video and audio analysis engine for multimodal content understanding.
- Developed an LLM Experimentation Platform and AI SDK for running and comparing LLM experiments.
- Stood up Google Cloud infrastructure — Compute Engine, Cloud SQL, Pub/Sub, BigQuery, Cloud Run, Cloud Build, and Secret Manager.
- Built internal tooling spanning CRM, alerting/monitoring servers, and GenAI-powered finance tooling.
- Implemented monitoring and observability for LLM workloads in production.
- Python
- TypeScript
- FastAPI
- LLMs
- Vertex AI
- Google Cloud
- BigQuery
- Pub/Sub
- Cloud Run
- React
Feb 2025 – PresentEntrepreneur in Residence
Antler · Jakarta, Indonesia
Selected for Antler Indonesia Cohort 8 (ID8) — one of Southeast Asia's most selective founder programs — to build and validate a startup from zero.
- Built and validated Intura, an AI Visibility & Brand Research platform, through the program.
- Worked alongside operators, angels, and investors across the Southeast Asian startup ecosystem.
May 2022 – Jun 2025Senior Machine Learning Engineer
Finaccel Teknologi Indonesia · Jakarta, Indonesia
Owned production ML systems and internal tooling — recommendation, forecasting, computer vision, and the data platform behind them — with a focus on measurable business impact and cost efficiency.
- Developed, maintained, and monitored ML applications and internal tools end to end.
- Optimized data pipelines and end-to-end ML workflows, cutting service processing cost.
- Improved alerting, monitoring, and architecture patterns for data services.
▸ AI Marketing Platform
- Built a web platform with React, Express, and TypeScript integrated with a Python service.
- Deployed generative AI for marketing copywriting and image banner generation.
▸ Customer Relationship Management (CRM)
- Partnered with data scientists on product recommendation, propensity modeling, dynamic pricing, and customer segmentation — lifting transaction rate ~5–15% and reducing churn ~4–6%.
- Optimized cron services with big-data approaches, cutting processing time ~300% on average.
- Built an automated A/B testing flow using Multi-Armed Bandit.
- Tracked and monitored 30+ AI/ML applications (CPU/memory, query, and infra usage).
▸ Operations, Data Platform & Internal Tools
- Built and monitored computer vision solutions — OCR and Face Anti-Spoofing — that reduced operational cost.
- Built a centralized 3rd-party data platform handling 300 requests/minute at <25% average CPU/memory.
- Built forecasting for Gross Marginal Value, run-out balance, and revenue — preventing ~4–8% of transaction failures and capping budget overruns at 20%.
- Python
- Scikit-learn
- LightGBM
- Prophet
- PyTorch
- OpenCV
- Kafka
- Airflow
- BigQuery
- FastAPI
- React
Oct 2021 – Apr 2022AI Engineering Analyst
Accenture · Jakarta, Indonesia
Migrated and modernized a large AI chatbot platform from on-premise to AWS, with Kubernetes and service discovery.
- Containerized chatbot services totalling 90+ microservices.
- Migrated 90+ microservices along with data storage and monitoring tools from on-premise to AWS.
- Revamped onto Kubernetes — cutting peak latency ~50%, reducing infra-related error rate from 9% to 6%, and handling 100–300 more requests/minute.
- Integrated service discovery with Consul for the AI chatbot services.
- Python
- Docker
- Kubernetes
- AWS
- ECS
- EKS
- Consul
Jan 2020 – Oct 2021AI Development Specialist
Bhinneka Mentari Dimensi · Jakarta, Indonesia
Modernized data and scraping infrastructure for a major Indonesian e-commerce platform and shipped recommendation pipelines.
- Re-platformed the scraping engine from on-premise to Apache Airflow (Cloud Composer), Cloud SQL, Dataproc, and BigQuery — cutting monthly infra cost by ~30M–100M IDR.
- Built Bulk Upload Tools to onboard new merchants, contributing ~1–3% of merchant acquisitions.
- Designed, built, and maintained data pipelines including content-based product recommendation and e-commerce scraping.
- Python
- Apache Airflow
- Cloud Composer
- Dataproc
- BigQuery
- Cloud SQL
Pub/Sub
Cloud Run
Scikit-learn
OpenCV
Dataproc
Cloud SQL// projects
Selected work
Production systems behind real business outcomes — recommendation, generative AI, forecasting, and the data platforms underneath.
Product Recommendation Engine
Personalized recommendations that moved revenue, not just metrics.
An end-to-end recommendation system combining content-based and propensity models with dynamic pricing and customer segmentation. Served through a high-throughput pipeline and continuously optimized with Multi-Armed Bandit A/B testing.
- Lifted transaction rate ~5–15% and reduced churn ~4–6%.
- Cut pipeline processing time ~300% with big-data optimization.
- Automated experimentation via Multi-Armed Bandit across live traffic.
- Python
- Scikit-learn
- LightGBM
- Kafka
- Airflow
- AWS
- BigQuery
- FastAPI
AI Visibility & Brand Research Platform
Measuring how brands show up inside AI answers.
A platform that evaluates brand visibility across AI assistants by generating realistic user queries, running them through multiple LLMs, and analyzing how (and whether) a brand appears — turning generative search into a measurable channel.
- Generative-search visibility scoring across multiple LLM providers.
- Query generation + response analysis pipeline at scale.
- Serving B2B clients across fashion, retail, SaaS, edutech, and beauty.
- Python
- LLMs
- OpenAI
- Claude
- Gemini
- Vertex AI
- BigQuery
- Pub/Sub
- Next.js
Social Media Listening Platform
Scraper pipelines + multimodal analysis across five networks.
A social listening product with dashboards and resilient scraper pipelines for TikTok, Instagram, Threads, and X (Twitter), plus a video and audio analysis engine for multimodal content understanding.
- Scraper pipelines for TikTok, Instagram, Threads, and X.
- Video + audio analysis engine for multimodal signals.
- Dashboards turning raw social data into client-facing insight.
- Python
- Pub/Sub
- Airflow
- BigQuery
- Google Cloud
- React
- TypeScript
LLM Experimentation Platform & AI SDK
Run, compare, and observe LLM experiments in one place.
An internal platform and accompanying AI SDK for designing LLM experiments, comparing prompts and models, and shipping production-grade observability and monitoring for LLM workloads.
- Unified experimentation across prompts and model providers.
- AI SDK to standardize LLM calls across services.
- Production observability for LLM cost, latency, and quality.
- Python
- FastAPI
- LLMs
- Vertex AI
- Cloud Run
- Monitoring
Financial Forecasting Suite
Forecasts that prevented failures and capped overruns.
Time-series forecasting for Gross Marginal Value, run-out balance, and revenue — feeding operational decisions that prevented transaction failures and kept budgets in check.
- Prevented ~4–8% of transaction failures.
- Capped budget overruns at 20%.
- Forecasting across GMV, run-out balance, and revenue.
- Python
- Prophet
- LightGBM
- Airflow
- BigQuery
Computer Vision for KYC
OCR, face anti-spoofing, and a KTP reader for identity flows.
Production computer-vision systems — OCR, Face Anti-Spoofing, and an Indonesian ID-card (KTP) reader — that automated identity verification and cut operational cost, built on metric-learning research.
- OCR + Face Anti-Spoofing reducing operational cost.
- Automated KYC via a KTP (ID card) reader.
- Grounded in published metric-learning research (ICACSIS 2020).
- Python
- OpenCV
- PyTorch
- CNN
- Metric Learning
Real-time 3rd-Party Data Platform
300 requests/minute at under 25% utilization.
A centralized platform for ingesting and serving third-party data in real time, engineered for high throughput and low resource footprint with strong monitoring and resilience patterns.
- Handled 300 requests/minute at <25% average CPU/memory.
- Centralized 3rd-party data access across services.
- Resilience patterns: rate limiting and circuit breakers.
- Python
- Kafka
- FastAPI
- Redis
- Kubernetes
- BigQuery
// courses
Free courses I've built
Hands-on data science courses distilled from the bootcamps I taught — interactive visualizations, in-browser Python, quizzes, and downloadable notebooks. Completely free.
// teaching
Sharing what I've learned
Years of teaching data science, ML engineering, and full-stack development — from bootcamps to a digital-bank big-data program.
Part TimeHead of Big Data & Machine Learning Instructor — Digital Bank
Rakamin Academy · Jakarta, Indonesia
Apr 2023 – May 2023
- Taught production big-data tooling: Spark, Hadoop, Kafka, and Airflow.
- Built end-to-end ML solutions for a chatbot.
- Built end-to-end ML for KYC automation, including a KTP (ID card) reader.
- Built ML monitoring and alerting for production environments.
- Spark
- Hadoop
- Kafka
- Airflow
Spark
Part TimeAI Engineering Instructor
Jakarta Labs · Jakarta, Indonesia
Apr 2022 – Jun 2023
- Taught ML deployment: serving models via REST APIs, building event-driven systems, and cron-based ML jobs.
- Helped students get production-ready with Docker, Kubernetes, Git, and cloud computing.
- Docker
- Kubernetes
- Git
- Cloud
Part TimeData Science Instructor
Rakamin Academy · Jakarta, Indonesia
Apr 2022 – Oct 2023
- Taught Python fundamentals and Git for collaboration.
- Authored and evaluated curriculum and videos for data processing — filtering, sorting, feature engineering, deduplication, and cleaning.
- Python
- Git
- Pandas
Pandas
Full TimeFull Stack Web Development Instructor
Hacktiv8 (Hacktivate Teknologi Indonesia) · Jakarta, Indonesia
Jun 2018 – Jun 2019
- Taught back-end with the Node.js stack — REST APIs, GraphQL, SQL, and NoSQL.
- Taught front-end for web and mobile — jQuery, Vue.js, React.js, and React Native.
- Node.js
- GraphQL
- React
- Vue.js
- React Native
Node.js
Vue.js// speaking & milestones
Talks, awards & moments
Conference talks, workshops, published research, and a few milestones from teaching and speaking along the way.
Best Session Presenter Award
12th ICACSIS 2020 · Jakarta
Recognized for the best session presentation at the International Conference on Advanced Computer Science and Information Systems.
View8+
years in production AI
Years building ML & data systems
LLM Routing in Production
BandungPy meetup
Google Cloud Startup Program
Accelerator session
Antler
Founder residency
Presenter — 12th ICACSIS 2020
Conference talk · Jakarta
Presented the thesis on Generalized Proxy Anchor Loss.
View
Fraud Detection Training
Bank Indonesia
30+
live ML services
AI/ML applications monitored in production
Python for Data Science
Tech in Asia
Guest Speaker
Hacktiv8 × Tech in Asia
Pair-based & Proxy-based Losses for Face Image Retrieval
ICACSIS 2020
Published research on metric-learning losses for seen and unseen face verification.
Data Science Bootcamp
Rakamin × OCBC
Data Science Class
Binar × Bukopin
3.91
University of Indonesia
M.Sc. Computer Science GPA
Intura Free Class
AI engineering workshop
Co-Founder & CTO at Intura
Since 2025
Leading engineering for an AI visibility & brand research platform.
Evolve
Tech community event
Fraud Analytics Workshop
Bank Indonesia
Servo Motor Control for a CNC Loading-Unloading Robot
QiR 2017
Published at the 15th Quality in Research conference (undergraduate thesis).
// education
Education
Master of Science, Computer Science
University of Indonesia · Depok, Indonesia
2019–2021 · Minor: Artificial Intelligence
- Thesis: Generalized Proxy Anchor Loss for Seen and Unseen Data Face Verification.
- Published: Pair-based and Proxy-based Losses for Face Image Retrieval — ICACSIS 2020.
Bachelor of Engineering, Electrical Engineering
University of Indonesia · Depok, Indonesia
2013–2017 · Minor: Control Systems
- Thesis: Modeling and Analysis of Servo Motor Control for a Loading-Unloading (LUL) Robot on a CNC Milling Machine.
- Published: The 15th Quality in Research (QiR), 2017.
// credentials
Awards & Certifications
Best Session Presenter Award — 12th ICACSIS 2020
ICACSIS · Jakarta · 2020
Presenter Participant — 12th ICACSIS 2020
ICACSIS · Jakarta · 2020
Data Visualization and Machine Learning
Algoritma Data Science School
Machine Learning Bootcamp
Evolve Machine Learners
JavaScript Full Stack Developer Bootcamp
Hacktiv8
// writing
From the blog
Field notes on AI engineering, ML systems, and the data platforms that run them.
Let's build something intelligent
Have a product to ship, a team to level up, or an AI problem that needs an engineer who can take it from data to deployment? I'd love to hear about it.