Data Scientist; Machine Learning

Position: Data Scientist (Machine Learning)

VentureSEA builds data and AI solutions for companies across Southeast Asia. As a Data Scientist you will work embedded with one of our clients and own machine learning projects from the first conversation to a model running in production.

Most of the work is traditional machine learning on tabular and time-series data
: gradient boosting, regression, classification, anomaly detection, forecasting and optimisation, usually on large and messy real-world datasets.

We are also starting to build agentic AI systems for our clients, so you should be familiar with LLMs and agent frameworks and comfortable picking them up as projects need them.

You will own your projects end to end. That means scoping them with the client's team, getting the data into shape, building and validating models, shipping them to production and making sure they keep working. You will be the client's day-to-day contact for your projects.

This is a fully remote, full-time role for candidates based in Indonesia.

What you will do

Turn messy operational data into reliable datasets. That data can come from sensors, machines, transactions or business systems, and it has gaps, outliers, duplicates and mislabelled records.

Build and validate models for problems such as anomaly and fault detection, predictive maintenance, forecasting, and optimisation and decision support.

Use time-based validation and honest backtests. Tell the client when the data cannot support a claim.

Take models to production:
Python services, Docker, scheduled training and inference, integration with the client's systems, monitoring and retraining.

Work directly with the client's engineers, operations teams and managers. Run check-ins, explain results in plain language, agree what success looks like, and chase down whatever is blocking progress, such as data access, unclear definitions or sign-off.

Write clear reports and documentation that non-technical stakeholders can act on.

Requirements Must have

At least 3 years of hands-on data science or ML engineering experience, delivering models that went into production or into real business decisions.

Strong traditional machine learning on tabular and time-series data: feature engineering, gradient boosting (XGBoost, LightGBM or similar), regression, anomaly detection, and leakage‑free validation over time.

Experience across the whole ML lifecycle: data pipelines, training, evaluation, deployment, monitoring and retraining.

Strong Python (pandas, Num Py, scikit‑learn) and SQL. Comfortable with Git, Docker and writing code other people can run.

Ownership. You drive a project forward without being chased, raise risks early and unblock yourself and others.

Clear written and spoken English, and the confidence to work directly with clients.

Familiarity with LLMs and agent frameworks, such as the OpenAI Agents SDK, Google ADK, Claude Agent SDK or Lang Graph: tool calling, prompting and evaluating agent output.

Nice to have

Experience with sensor, IoT or telemetry data, for example in manufacturing, energy, utilities, logistics or facilities.

Physics‑informed or grey‑box modelling, or a degree in engineering or another quantitative field.

Workflow and serving tools such as Airflow or Prefect, FastAPI, Postgres or SQL Server, and deploying to client infrastructure, including on‑premise environments.

Consulting or client‑facing project experience.

This role is probably not for you if your ML experience is mainly with LLMs, RAG or chatbots, and you have not built and shipped traditional ML models on structured or time‑series data.