We love building software for the web — with over a decade of experience in building software, we have backgrounds in software engineering, product management and technical leadership.
Individually, we’ve created or helped create dozens of software products from the ground up. Together, we’ve started Subthread: a product studio to help companies big and small go from idea to product.
We created multiple AI-powered apps and prototypes to help OpenUp experiment with the value of AI for their business:
We’re very excited to keep helping OpenUp execute on their ideas in the fast-changing landscape of AI.
Generate an image of me standing next to the Egyptian pyramids, wearing a bathrobe. It is the middle of the night and the moon shines bright.
For a large TelCo we built a custom AI backend that would take a headshot / image of a person and create a new image in a totally new context, in a completely new style. We accomplished this without the need to train a new model, saving money and time.
As part of the AI pipeline we also implemented a speech-to-text LLM so that users who would use the product could instruct the AI what type of image they want.
They used this backend as a part of a marketing event where people could walk into a booth and have their “AI picture” taken. So far our backend has created thousands of pictures.
Quest was a business-to-consumer podcasting platform designed to cater to both podcast authors and listeners. While the platform was ultimately sunsetted, our contributions played a pivotal role in testing the product-market fit that Quest's co-founders were seeking.
We achieved this by rapidly implementing new ideas, offering valuable technical insights, and transforming concepts into fully-fledged products within a matter of weeks.
One of the first projects we worked on together, Dagstaat is a full-featured SaaS application that supports:
We’re proud to have been able to take our client’s idea to production as well as continuously expanding and maintaining the product in the 3 years since launch.
We boosted the recruitment pipelines of a popular recruitment company using AI, which allowed them to save up to 2000 hours of work every year.
Using powerful AI methods, we scrape, parse and convert countless resumes so that recruiters can quickly see whether or not a candidate is a good match for a job. If the candidate is a match, a resume file is automatically generated that matches the recruitment company’s template and tone of voice so all there’s left to do is share the generated resume with their customer.
For a company that publishes academic research papers, we used a ‘traditional’ machine learning approach to automatically classify 20.000 customer support emails. We identified several clusters that each email would belong too, and then used Generative AI to come up with a tag or label to name the cluster.
The outcome was that the customer got some great insights into the frequency of customer support requests, which allowed them to prioritise where to put time and money to further optimise their customer support department.
While these two companies have distinct profiles, the nature of our work with them is similar: a technical overhaul of outdated codebases.
For uButler, we played a key role in transitioning their Meteor-based code to React. Similarly, our collaboration with GoodUp involved the replacement of their outdated Ember application with a brand-new React-based solution.
In both instances, our contributions accelerated the code rewriting process, while guiding and teaching our clients’ internal tech teams about the latest best practices when it comes to frontend application development.
We design and build software for the web.
We help companies big and small to explore new ideas, experiment with new technologies or even start new businesses entirely.
Here’s some examples of things we have built in the past: