Projects

We pride ourselves on hiring great engineers from a variety of disciplines. As such, we are able to deploy a broad set skills for our clients. Since a list of all of the things we do might become tedious, we opted to provide a filterable list of recent projects.

Ensure that villages in Africa have access to fresh water

IBM invents things that matter to the world.

Device Service Do good

Situation
Aid agencies spend significant capital deploying water wells and other improvements in the developing world. Half of the installed equipment breaks down within 18 months, and can take up to 7 months to repair or replace.

Task
Provide rural infrastructure with a satellite connection so repair can be efficiently coordinated.

Action
Geocene worked with IBM and SweetSense to build a rugged solar-powered and satellite-connected feedback system.

Result
A system that monitors and protects the drinking water of over 2 million people.

Reflection
Satellite data is highly reliable but very expensive. Disrupting competition is on the horizon.

Hardware
  • Injection Molding
  • Parametric modeling
  • PCB Design
  • FreeRTOS
  • Battery Power
  • Solar Power
Communications
  • Irridium Satellite
  • Swarm Satellite
Backend
  • AWS
  • S3
Data-science
  • Sisense
  • BI Dashboard

The Geocene Engine in an RMS current logger

The World Bank is an international financial institution that provides loans and grants to the governments of poorer countries for the purpose of pursuing capital projects.

Owned IP Device Do good

Situation
Time of use data requires eyes-on monitoring of power meters, a time consuming activity.

Task
Automate collection of power data in areas without reliable connectivity.

Action
Geocene repurposed the Geocene Engine used in the temperature data logger to collect Root Mean Square (RMS) data.

Result
Multiple funded research projects.

Reflection
The Geocene Engine is a powerful tool, useful for logging any number of envorinmental variables.

Hardware
  • Injection Molding
  • Parametric modeling
  • PCB Design
  • FreeRTOS
  • Battery Power
Communications
  • Bluetooth
Mobile
  • Android
  • iOS
  • Kotlin
  • SWIFT
  • OKHTTP
  • Native BLE
  • Native code
  • Core Data
Backend
  • RESTful
  • S3
  • Django
  • Collect
  • Postgres

Embed with a startup to accelerate development

Arable Labs builds affordable tools that help people collect and synthesize site-specific agricultural data.

Existing team Device

Situation
Arable needed a little extra help to develop the features most commonly requested by their clients.

Task
Build a suite of new features and functionality while providing engineering management and strategic guidance.

Action
Geocene got right to work and deployed multiple engineers on a full time basis, relieving pressure on a skilled existing team.

Result
A successful B fund raise during the COVID-19 epidemic.

Reflection
We love a challenge, whether it is building something from scratch or integrating directly with an existing team.

Hardware
  • Solar Power
  • Battery Power
  • FreeRTOS
  • C
  • Compression
Communications
  • NB IoT
  • LTE M
Backend
  • RESTful
  • S3
  • Django
  • Postgres
Data-science
  • Machine learning
  • Edge computing

Easy-to-install whole building load monitoring

Persistent Efficiency produces and commercializes sustainable energy-monitoring solutions.

Device Early stage

Situation
Power monitoring is essential for building managers, but installation is costly and requires downtime.

Task
Develop a stick-on power patch capable of analyzing a building's energy consumption

Action
Geocene worked with Persistent Efficieny's engineering team to develop the power patch.

Result
Installed functioning patches in dozens of commercial buildings throughout the Bay Area.

Reflection
Building energy monitoring is critical for an energy efficient future facing climate change.

Hardware
  • FreeRTOS
  • Embedded Linux
  • Digital Signal Processing
  • Custom Protocol
  • Injection Molding
  • Custom Calibration
  • Custom Test Fixture
  • Parametric Modeling
Communications
  • WiFi
  • Bluetooth
Backend
  • RESTful
  • S3
  • Django
  • Collect
  • Postgres

Detailed power consumption data logging

Sustainabilist enables higher profit and lower risk in cleantech with scientific process improvement, custom software, data science, and machine learning.

Owned IP Device Do good

Situation
Data science and machine learning tools function best with detailed input data.

Task
Create a cost-effective data logger that produces detailed power consumption information.

Action
Geocene repurposed the Geocene Engine used in the temperature data logger to collect detailed power consumption data.

Result
Enabled our client to achieve a 16% reduction in costs.

Reflection
The Geocene Engine is a powerful tool, useful for logging any number of envorinmental variables.

Hardware
  • Injection Molding
  • Parametric modeling
  • PCB Design
  • FreeRTOS
  • Battery Power
Communications
  • Bluetooth
Mobile
  • Android
  • iOS
  • Kotlin
  • SWIFT
  • OKHTTP
  • Native BLE
  • Native code
  • Core Data
Backend
  • RESTful
  • S3
  • Django
  • Collect
  • Postgres

An easy-to-use ML platform for detecting fire

The National Institute of Health is the nation’s medical research agency — making important discoveries that improve health and save lives.

Owned IP SUMS Service Do good

Situation
The National Institute of Health wanted to know which people were making cooking fires and what stoves they were using. They had a lot of temperature data, but but it wasn't organized to make it useful.

Task
Build a tool that allows non-technical clients to use machine learning on their own data.

Action
Geocene built a cloud ML platform that automatically analyzes client data for cooking events and allows for sorting and filtering on user tags.

Result
Over 1 million cooking events identified.

Reflection
Analyzing data isn't useful only at the end of the project. Live analysis can catch problems early and give users a chance to fix them before they become catastrophic.

Backend
  • S3
  • Postgres
Data-science
  • R
  • Periscope

A new approach to remote data collection

The Clean Cooking Alliance works with a global network of partners to build an inclusive industry that makes clean cooking accessible to the three billion people who live each day without it.

Owned IP Device Do good

Situation
Cookstove usage in very remote areas was inconsistent and data about usage was expensive to collect.

Task
Computerize distributed data from remote locations that have little or no internet connectivity.

Action
Geocene designed and built a modern sensor and data logger that connects to a custom mobile app – users can use their phones to carry the data from villages to the city to upload.

Result
10,000 sensors deployed and over 1 billion data points collected.

Reflection
It isn't just the developing world with poor access to the Internet. Similar solutions might be useful in buildings under construction, sub-basements, and undeveloped land.

Hardware
  • Injection Molding
  • Parametric modeling
  • PCB Design
  • FreeRTOS
  • Battery Power
Communications
  • Bluetooth
Mobile
  • Android
  • iOS
  • Kotlin
  • SWIFT
  • OKHTTP
  • Native BLE
  • Native code
  • Core Data
Backend
  • RESTful
  • S3
  • Django
  • Collect
  • Postgres

Capture environmental data with Google Cars

The EDF addresses today's most urgent environmental challenges.

Service Do good

Situation
Google Street View Cars drive everywhere and collect lots of data, but it is disorganized and hard to use.

Task
Create a data pipeline for the massive quantity of air quality data collected by Google Street View cars

Action
Geocene built an ingress, cleaning, parsing, and data warehousing solution using Google Cloud Platform to empower Environmental Defense Fund’s data science team to analyze these data.

Result
Our client moved from a tangle of spreadsheets, CSVs, databases, and flat files to being able to run simple queries using BigQuery to access the data that mattered most to them.

Reflection
Organizing data empowered a data science team to perform their best thanks to Geocene's engineering.

Backend
  • Google Cloud Platform (GCP)
Data-science
  • BigQuery
  • Google Maps
  • YAML
  • Python
  • Pandas

3D greenhouse data collection

Reliable Addressable Devices identifies and eliminate the microclimates in greenhouses.

Early stage Device

Situation
Most indoor and greenhouse cultivation facilities use just a couple of sensors to cover hundreds of thousands of square feet. This leads to delays, substandard crops, pests, and wasted energy.

Task
Create strands of inexpensive sensors connected via TCP/IP standard network that could be hung in greenhouses.

Action
Geocene created sensor "nodes" that not only capture data, but relay information to users using LEDs and Augmented Reality.

Result
Client is pursuing funding with a functional prototype.

Reflection
Presenting 3D data in 3D space provides powerful access to data without needing to sit in front of a computer.

Hardware
  • Vacuum Forming
  • Parametric modeling
  • PCB Design
  • FreeRTOS
  • Power over Ethernet
  • LEDs
Communications
  • TCP/IP
  • RS232
Mobile
  • iOS
  • Augmented Reality
  • ARKit
  • Vuforia

Asset tracking for remote areas

UCLA's primary purpose is the creation, dissemination, preservation and application of knowledge for the betterment of our global society..

Device Do good

Situation
Laterns and other development assets distributed in the developing world are frequently stolen, sold, or otherwise misappropriated.

Task
Create a non-obtrusive way to identify if assets are where they are supposed to be without using

Action
Geocene created Bluetooth beacons, a smartphone app, and a backend capable of using Bluetooth signals to log and plot device locations

Result
Beacons and app have been distributed, asset project awaiting deployment.

Reflection
Relatively simple solutions to significant development problems can go a long way, with a little effort into applications.

Communications
  • Bluetooth
Mobile
  • Android
Backend
  • S3
  • Postgres
  • Django
  • API

Make it easy to train sets of data for machine learning

The National Institute of Health is the nation’s medical research agency — making important discoveries that improve health and save lives.

Open source Service Do good

Situation
Machine learning is powerful, but requires training the machine. Manually labelling training data is cumbersome.

Task
Make it easy for people without training to label training data.

Action
Geocene created an open-source graphical tool that allows users to quickly select the key parts of their data for machine learning.

Result
A widely used web platform.

Reflection
Open-source and machine learning tools are vastly more powerful when they are accessible.

Backend
  • vue.js

Turn a consumer product dream into a reality

Some clients have asked us to stay quiet about our relationship. In keeping with their request for anonymity, we only list here the tools we used to build their products.

Device Private Early stage

Situation
A private client had an idea for a product to win early funding.

Task
Work with the private client to turn an idea into a consumer product.

Action
Spec'd, designed, and built dozens of iterative functional prototypes.

Result
Client successfully raised more than $10 million in their next funding round.

Reflection
A great product look and feel helps build investor confidence.

Hardware
  • CNC Machining
  • Parametric modeling
  • PCB Design
  • FreeRTOS
  • LEDs
Communications
  • Bluetooth
  • Custom Protocol

An easy-to-use ML platform for detecting broken equipment

The University of Colorado Boulder is all about realizing the positive impacts of new knowledge.

Service Do good

Situation
CU Boulder has collected a significant amount of data about equipment functionality in distant parts of the world.

Task
Empower users to analyze their own equipment functionality data.

Action
Geocene built a cloud ML platform that automatically analyzes client data for equipment breakdowns and issues alerts to users.

Result
Tens of thousands of equipment failures identified.

Reflection
Machine learning can help efficiently deploy maintenance staff.

Backend
  • S3
  • Postgres
  • API
Data-science
  • R
  • Periscope