PH - Machine Learning Researcher, Geospatial

Build next-level machine learning products for geospatial use cases in our collaborative team...

Thinking Machines is a technology consultancy building AI & data platforms to solve high impact problems for our clients. Our vision is for the Philippines to become a global hub for data science. To do that, we create data cultures, one organization at a time.

We’re a company made up of intellectually curious, civic-minded, forever-learning individuals. We believe that great data science products are built with care for people, and that the best way to drive inclusive innovation is to start with a diverse team.

Our field of work is incredibly dynamic, so we want to work with people who are committed to growing with us. We want to hire people who can demonstrate an ability to learn, then provide them with personalized coaching, growth opportunities, and a great working environment to get them to world-class.

If you have a portfolio of work in machine learning and are interested in developing your skills in the exciting field of geospatial data science, this is an opportunity for you!

Our Geospatial Machine Learning Team customizes AI models to extract useful, actionable insights from satellite imagery and other earth observation data sources.  We frequently work with organizations in the development and social impact sectors (our clients include UNICEF Innovation Fund, United Nations Development Program (UNDP), Asian Development Bank, and World Bank, and iMMAP), but we’ve also developed geospatial applications for private sector clients including telecommunications and real estate companies. Our past work includes:

• Mapping Informal Settlements in Colombia with Machine Learning and Low-Resolution Satellite Imagery. We worked with the humanitarian agency iMMAP in Colombia to use AI to help the locations of at least 400 informal settlements occupied by migrants from Venezuela. 

• Interpretable Poverty Mapping using Social Media Data, Satellite Images, and Geospatial Information. Using a combination of data from Facebook’s marketing API, low-resolution satellite imagery, and OpenStreetMap, our team developed an AI model that could estimate average household wealth for every 18 km2 of the Philippines. The model was developed with support from the United Nations Development Program (UNDP) and in collaboration with the Zero Extreme Poverty Movement.

• Predicting Property Values for a Real-Estate Digital Platform. Our team worked with OhMyHome, Singapore’s leading property transaction ecosystem, to build a custom machine learning model that can estimate the value of any real estate property in Singapore and Malaysia within 4% difference from the actual price.

• Supporting broadband infrastructure planning for Globe Telecom. We worked with Globe Telecom to build an Artificial Intelligence model that detects the location and measures the socioeconomic class (SEC) of households across the entire Philippines. Using a combination of satellite imagery, telco usage data, and external location information, we were able to generate SEC classification estimates for each populated 50m x 50m tile in the Philippines.

This is an opportunity for someone with a strong portfolio of work in machine learning and geospatial analytics to do great work in the company of a high-performance team. This job demands creativity, critical thinking, and a focus on delivering excellent work. You’ll be expected to effectively handle projects from day one, and will be provided with effective guidance as you learn and implement machine learning methods for our clients and internal products.

We’re a growing tech consulting company with constantly evolving responsibilities, and the following is an incomplete but representative list of things you can expect to be doing:

• Developing machine learning models: feature engineering, experimenting with different model types, choosing the right metrics and measuring performance, hyperparameter tuning, feature importance, etc. Our Geospatial Team works a lot with computer vision in particular.
• Exploring geospatial data and its sources: looking for patterns and signals in the data, identifying important fields, gleaning insights, and discerning whether machine learning can even apply to the problem. Expect to handle a lot of remote sensing data and satellite imagery using geospatial data processing tools like GeoPandas, QGIS, and BigQuery GIS.
• Understanding model design and decisions: explaining how models work under the hood and how a model arrived at a certain prediction, choosing the right model family, and justifying accuracy tradeoffs for faster performance
• Research and keeping up with spatial data science space: You’ll need to keep up to speed on the latest research at the intersection of remote sensing, spatial analytics, and machine learning by reading through published ML papers, and implementing improved versions of their models in new domains.
• Acquiring, cleaning, and evaluating the integrity of datasets! Let’s be real, this is a tricky and satisfying 25% of the job. We use Google Cloud Platform for many components of our machine learning workflow.

We’re looking for someone who meets the following profile:

• Comfort with machine learning models - You have a demonstrated understanding of classic machine learning models for classification and regression (e.g. decision trees, ensemble models, SVMs etc.).

• Comfort with code - You can use your local machine to scrape, load, and parse through moderately large datasets without much handholding.

• Clear communication - We help clients get the most out of their data, so we diagnose their needs effectively, do the analysis correctly, and communicate our findings in a way that leads to understanding and action. At minimum, you need to have the ability to articulate your points logically, and have to be willing to learn this skillset as you grow with us.

• Productive curiosity - You ask a lot of the right questions. Find a surprising correlation? Dig into the raw data to validate it.

• Enjoys both teaching and learning - We believe that data science is an incredible field to be in today. There’s a huge amount of new material to learn, we all want to learn it, and we’re looking for someone who wants to contribute to everyone’s growth. As part of our job interview, we’ll be asking you to read and summarize a machine learning paper for us.

• Strong sense of initiative—You’re always looking for ways to be useful and you hate having nothing to do.

• Social intelligence— It’s extremely important that you work well with others and thrive in an environment with lots of teamwork and interpersonal interaction.

• You have a portfolio of at least two projects that you’ve done involving predictive analytics or machine learning
• Undergraduate/Graduate degrees in Computer Science, Physics, Mathematics, Statistics, or any related field, OR relevant work experience
• Strong fundamental statistics skills, and linear algebra handling skills
• Strong quantitative analysis skills
• Familiarity with statistical programming languages like Python, or R

While you don't need to have previous experience working with geospatial data, we are looking for candidates who are willing and eager to learn:

• Working on geospatial workflows wrangling geospatial data types. This includes extracting data from raster files, transforming and performing spatial operations on location data and visualizing data on a map.
• Using geospatial tools for processing, visualization and analytics such as GeoPandas, QGIS, BQ GIS, CARTO, Mapbox, etc. 
• Training Computer Vision models, preferably with georeferenced imagery such as satellite or drone imagery.

We offer the following compensation and benefits:

• Competitive salary — the compensation amount is positively correlated with the difficulty and scope of the job, relevant experience, fit, and skill factors.

• Fully remote — due to the global pandemic, we have shifted to a fully remote company for the foreseeable future while we monitor the situation.

• Individual professional development budget — an annual budget for conferences, training courses, books, and software are available to sharpen your skills and build new ones to help you grow in your role.

• Full health benefits — generous health insurance package, including sick leaves and a health card immediately upon hiring.

• Regular 1:1 meetings with the leadership team to discuss career and personal goals, job progress and any questions and concerns.