The first stage consists of general questions on experience and fit. The second stage was an online technical test on software engineering / reinforcement learning / machine learning. The final stage was another general interview but with more members of the team.
3rd -
Interview questions [1]
Question 1
They asked about my experience to grasp my fit to the position
I applied online. The process took 1 week. I interviewed at AgileRL (London, England) in Sep 2023
Interview
An initial chat with the CEO, followed by a four-part technical assessment and lastly a chat with a co-founder. The product they're working on seems really interesting and genuinely useful to the open-source RL ecosystem. The CEO was great to chat to and learn about. The technical assessment is with TestGorilla and had four parts. 1: Dynamic Programming with Lists in Python, 2: Data Structures and Algorithms in Python, 3: Reinforcement Learning 'Test' and 4. Deep Learning 'Test'. On the technical side I fell short on the tests - they were very heavily academically aimed with lots of multiple-choice questions on curveball/trivia from some 1980's reinforcement learning techniques and a pen-and-paper maths focused deep learning test. Most of the nomenclature and techniques of these tests are something academically tested by rarely used by industry or common frameworks. I withdrew my application after crashing the 'tests' as either the tests didn't cover the actual activities of the role, or I was woefully unsuitable for the role.
Interview questions [1]
Question 1
Something like 'Temporal Difference Difference learning is implemented to solve a problem with the '[Parameter] being changed, what are the other two parameters that might be changed?'