Algorithm Architect



Don't Apply If:

  • You care more about winning an argument than finding out the truth
  • You thrive in office politics
  • You think being polite is more important than being honest
  • You care more about sounding smart than being understood
  • You get annoyed when you are asked to work on a problem that’s outside of your typical job description and that you may fail at
  • You don’t think you should be expected to work on problems outside of your comfort zone if you “do your job”
  • You think your manager should give you clear instructions on how a problem can be solved and/or where to look for a solution
  • You have deep trust in established authority

Do Apply If:

  • You love being pushed to do things you didn't think you’re capable of
  • You are excited to work on hard problems, especially if others already failed trying to crack them
  • You hate being told what to do
  • You communicate clearly and use precise language and insist that others do, too
  • You find joy in finding out how things work
  • You can’t keep quiet if you think something doesn’t make sense
  • Not knowing if you are correct bugs you, even if nobody else noticed
  • You don’t beat around the bush; you say what you really think, even if it doesn’t sound nice or polite
  • You are a critical thinker

What We're Working On

We are building an influence algorithm. In other words, we are trying to find ways to describe groups of people mathematically. Many tried and failed before. But we think we can make it work.

Our core hypothesis is that influence can be quantified by tracking attention flows. In order to do that, we ingest data streams from multiple sources (we started with Twitter and are now indexing podcasts and soon more). We then cross-reference these datasets in an attempt to continuously improve the accuracy.

The accuracy of our work is being verified by members of the groups that we aim to describe. We publish our results in real-time and there are thousands of people already using our scores. It is hard to verify when we are right. But it is very easy to tell when we are wrong. This short feedback loop puts us in a unique position to work on problems that might be much harder or impossible to solve somewhere else.


Why - our Mission

We are working on decentralizing information distribution.

Distribution of information is controlled by those who are in charge of:

  1. Broadcasting channels
  2. Credibility signaling

Society suffers when small groups control what others get to know and believe. Be it through controlling the broadcasting (tv, radio, newspapers, access to social media platforms) or credibility signaling (who should be trusted and who shouldn’t).


Work setup

We are a small, VC-funded startup. We are a remote-first team. Most of the team is based in Europe (Germany, UK, Spain). You can make your own hours, but everybody is expected to be online during office hours in CET. We try to meet in person and work together for several days at least every 3 months. Other than that the company ‘lives’ in Slack, Notion, Tandem and other tools enabling effective communication.


About this role

You will be working on the core algorithm. This means that you will look for creative, but methodically robust ways of identifying groups in various streams of data (e.g. Twitter, Podcast RSS feeds, meetups, Reddit, GitHub etc.). You will design and execute experiments closely following the scientific method.

You should be able to design and perform experiments independently. We have a dev team that’s responsible for engineering of the data streams. You need to be able to code up experiments from A to Z and give clear instructions to the devs what data streams you need.

You will work directly with the founder, who is also working on the algorithm. You will be responsible for implementing the new algorithms and will be taking the lead in shipping these into production along with our dev team. You will make sure that all changes are properly documented. You will also be responsible for making sure that there are robust tests put in place to ensure data correctness. You will often communicate with engineers on our dev team and you will be able to request resources from this team.

You will have plenty of flexibility and you will be encouraged to try new things and think outside of the box. The only requirement is that you have to be able to defend the logic behind these ideas and provide a falsifiable methodology.


Responsibilities

  • Designing & running experiments for improving the algorithm
    • Testing to verify performance of these experiments
  • Identifying issues with the algorithm (e.g. monitoring for bugs, inaccurate results, false data inputs/outputs etc.)
    • Proposing potential solutions and methods for falsification
  • Communicating with the dev team
    • Making sure the algorithm is being implemented correctly
    • Making sure that the tests for data accuracy are sufficient and are running correctly
  • Writing documentation for the approved algorithm, but also for all experiments (both failed and successful)
  • (Optional) Building simple machine learning models

Requirements

  • Good grasp on statistics
  • Good grasp on Probability Theory
  • Coding skills (preferably R and Python)
    • Django experience would be good for database interaction
  • Experience working with stats programs (e.g. SPSS or R)
  • Mathematical thinking
  • Relevant working experience

Great to have

The following interests are not required, but if you share some of them, you will find this role more exciting. You will be encouraged to explore these interests further and look for ways to bring them to your work and research.

  • Data engineering experience (working with Data APIs and Relational Databases, data cleaning and transformation)
  • Interest in Machine Learning; especially if you can build & train simple models
  • Familiarity and interest in Memetics, Information Theory, Graph Theory, Network Theory
  • Familiarity and understanding of Bitcoin network (especially the role of Proof of Work)
  • Familiarity and interest in Deep Learning

Compensation

€45k – €75k per year

0.15% – 0.3% Equity


How To Apply

Please answer our 5 questions and apply here