Geli

Geli is accelerating the transition to clean energy by connecting and optimizing storage and renewable resources through world-class software and solutions

26 - 50 employees

11 - 25 engineers

$10m - $25m funding

Series b

Apply

Geli provides software and business solutions to design, automate, and manage energy storage systems. Geli’s suite of products creates an ecosystem where project developers, OEMs, financiers, and project operators can deploy advanced energy projects using a seamless hardware-agnostic platform. Geli is a Series B company whose software actively manages megawatts of projects that are deployed around the world.

Why join us?

Work for a meaningful, rewarding cause — Fight climate change by enabling the integration of fluctuating renewable energy sources into the grid and increase their value through smart energy storage solutions.
Autonomously work on exciting challenges — We have no shortage of challenging problems in data science / machine-learning, IoT as well as data streaming and need more brains to help solve them.
Join the ride —We are just closing our Series B and are expanding quickly!

Engineering at Geli

Engineering team and processes

We work in two-week sprints with teams focusing on data science, our on-prem software, and the cloud-based services. We try to knock out well-defined epics across teams that complete in system integration testing. Our feature delivery process involves developing in our test environment, SIT on stage, and deploying to production. All stories undergo a PR review with a focus on clean, well tested and documented code. Our development tools include Bitbucket, Git, Jira, and Jenkins.

Regarding the organizational structure, our VP of Engineering sets our high-level goals, the product manager helps prioritizes business needs and scopes epics, and tech leads help supporting members of their team with design and guidance.

Technical Challenges

Over the last couple of months, we transitioned from running our algorithms on-premises to shifting the compute-heavy workloads to the cloud running on Kubernetes and making heavy use of messaging with RabbitMQ. Along with this development, there has been a rapid proliferation of micro-services such as a service that provides historical and forecasted weather data to our nodes, and a market connector that fetches wholesale electricity prices from various markets and sends that to our optimizers. We are in the early stages, and there are plenty of cloud infrastructure challenges ahead.

On the algorithmic side, the work done by the cloud infrastructure team has provided the opportunity to utilize more resource intensive algorithms. We are transitioning away from deterministic to stochastic optimization. This is not only an intriguing optimization problem but also an exciting forecasting challenge as we need to forecast distributions and generate multiple scenarios that accurately encapsulate the cost distribution.

Projects you might work on

Centralize the forecasting and optimization algorithms in the cloud! Currently, there are long-running processes that shadow each physical site. We want to move to a model in which a central forecaster and optimizer solves jobs for all sites which allows us to use a better, more expensive solver, is easier to maintain and provides better resource utilization.
Improve our simulation infrastructure! Presently, we use Airflow with ECS workers but would like to transition to running it all on EKS for better performance and resource utilization. Additionally, we would like to gain more insights into the data which requires storing more meta-data and deriving value through reports and dashboards.
Scale our data pipeline capabilities! We are constantly adding more and ever larger sites to our platform which our IoT device telemetry collection, processing, and persistence need to keep up with. This includes evaluating different messaging protocols (e.g. Kafka) and time-series databases as well as improving the ingestion, processing, and retrieval of telemetry (e.g. reactive streams).
Improve our probabilistic forecasts! Our ability to save money for our customers is a direct consequence of the load and solar generation forecasting accuracy. We are always innovating in the field by prototyping novel algorithms.
Build our Virtual Power Plant offering! So far, our main focus has been on optimizing behind the meter energy management for individual sites. There are more interesting use-cases coming up that require aggregating multiple sites and biding their capacity into marketplaces and utility programs like demand response.

Tech stack

Python

JavaScript

Java

Django

RabbitMQ

Kubernetes

Cassandra

Airflow

Docker

Kotlin

Spring Boot

Spring Cloud

TimescaleDB

Working at Geli

We are a fun group brought together by our cause to have a positive impact on the environment. Our nimble organization is rooted in meritocratic and autonomy, has a flat hierarchy with diverse, supportive coworkers who pull together across teams.

Perks & benefits