Photomath

See other positions

What do we do?

Photomath is changing the way students learn math. We’ve achieved viral, organic growth to become the most popular education app in the US with over 200 million downloads. With our mobile app, users can solve math problems instantly using their phone camera to get detailed step-by-step instructions to help maximize learning. Since 2016, Photomath has grown from 33 to over 200 people, with offices in Zagreb, Croatia and the San Francisco Bay Area. We are in the process of forming a new AI team and are looking for students who want to learn with hands-on AI work on many challenging AI/ML projects that we want to explore.

Role of the AI team is to solve seemingly unsolvable problems using cutting edge technology, especially focused on the solutions based on the fast changing and always stunning field of deep learning.

What is the Role?

Participate in the iterative AI problem-solving workflow:

1. cooperate in defining the problem, metrics, and brainstorming about potential solutions
2. engage in defining the data source and annotation process (we are using a dedicated annotation team)
3. develop a proof of concept
4. integrate a solution as a server-side or on-device service
5. architecture search optimize for quality in terms of accuracy and resource consumption
6. retrain the model on a constantly growing dataset
7. jump to 4.

Who are You?

• have an understanding of basic machine learning concepts taught at university courses or equivalent online courses
• have trained and evaluated a neural network using TensorFlow, Keras, or PyTorch
• have created a simple web or mobile app serving a simple neural network model (if you haven’t, feel free to try and send us the result :) - i.e. an idea: you can try to create a simplistic Photomath that adds up the numbers on the uploaded image)
• linear algebra and statistics pose no problems for you
• a desire to work on the whole interdisciplinary AI process from the idea to the actual implementation of the solution

A plus is if you have:

• basic knowledge of classic computer vision concepts and/or text analysis
• basic knowledge of deep learning models for computer vision and/or text analysis problems
• enthusiasm about some of the following: optimization, algorithms, c++, infrastructure, backend, whiteboard brainstorming sessions, computing with 5D tensors

What do we offer:

• intensive mentoring from our senior team members
• participation in whole-team per project brainstorming
• a stimulating environment always open for new ideas, potential problems, and potential solutions
• hundreds of terabytes of constantly growing data with an awesome supporting annotation team
• helping educate hundreds of millions of users worldwide