Our goal is to build a global invoice finance company that enables small businesses to instantly receive early payments against their outstanding invoices.
Results-oriented organization with zero bullshit
NoviCap is a meritocracy where ideas flow freely and are judged on merit. Management is very down to earth and frankly we are still a small company (<25 people) so you can have a huge impact.
We care about making our customers happy and build a great company together. In order to make sure we all agree on what is best for the company, we are very transparent with our numbers and regularly present board decks to the entire company. We also share meeting notes of the weekly management meeting. We have dashboards hanging in the office that display our key metrics so everybody knows how the business is doing at all times.
If you want to work in an environment with zero politics and where data carries the day, then you will love this.
We try to apply the best practices as they are known today in all fields, ranging from marketing to engineering. People learn new things every day and we are always open to trying new things. We have a big NoviCap library with many highly rated books on specific topics are present. If you are hungry to learn about business, risk, marketing, sales or scaling startups then you are going to love this.
We expect you to think critically, learn independently and figure things out.
If you are looking for a position with a lot of hand holding then please look elsewhere.
There is no shortage of technical challenges at NoviCap ranging from managing large datasets, complex trading algorithms, responsive UI, machine learning, pricing algorithms, API's (loads of them) and ofcourse keeping everything nice and tidy for maintainability.
We build major new functionality every month and currently these are some of the projects that we are ahead of us in the near term:
- Automated payments
- Implement direct debit infrastructure
- execute collection with the bank
- execute direct debit mandate
- track and execute recurring collections
- Reporting Implement US-based data sources and bank integrations
- Automate fraud analysis Invoice finance is a prime target for fraudsters and we have robust internal processes to detect it, these can be automated.
- Internal fraud tools
- audit trail
- maker/checker for sensitive actions
- backoffice permission system
- Iterate and continuously expand our pricing algorithm Our goal is to make sure we can price any invoice in the world instantly, currently our coverage is limited to Spain and UK.
We're using all the good stuff ;)
Below is an overview of used technology of our main projects and a short explanation of what the project is for:
- red shift
- AWS lamda for synchronization
- postgresql views
We value data-driven decision making a lot, no excuses or decisions made on feelings. We use amazon redshift to store daily snapshots upto 1 year back, this day-by-day view allows us to answer any question about the present and the past rapidly.
We use AWS lambda to load database backups into redshift and the user interface is ensured with metabase.
- Ruby on rails
Our web application is fully tested with more than 390 functional scenarios and 838 unit tests. The application is built on SOLID principles and we are using react.
- command line tools
We are very good at scraping and have amassed millions of records from social media data to company directory data. Our tool of choice is nodejs and elasticsearch for parsed storage.
We experimented with hadoop for processing but this turned out to be a bad idea.
- yeoman tooling
Our marketing website is fully static and therefore is impossible to hack, the website is precompiled with yeoman and uploaded to S3 and then delivered via cloudfront. Parts that require updating often are built with markdown.
Well, nobody has built a better CMS so far. Our wordpress is hosted behind cloudfront so any possible security issue (of which there are many) doesn't impact us.
Risk algorithm microservice
- random forests
- logistic regression
- lots of scraped data!
- scikit learn
We've built an machine learning algorithm that can predict with 86% accuracy if a company will default next year or not.
- heroku for deployment
- codeship for continuous testing
- git for version control
- github for code review
- hipchat for communication
- rollbar for error tracking
We are using Kanban to manage work in progress.
Official website, founding date, employees, how did it all begin... Do you know the whole story?Tell Us!