A genuinely high algorithmic bar, plus probability and puzzle rounds that most software loops never ask.
D. E. Shaw runs one of the harder algorithmic bars in the market, and it does not stop at data structures. Expect probability and combinatorics questions, brain-teaser style puzzles with a rigorous expected answer, and follow-ups that push on the mathematical reasoning rather than the implementation.
The systems side is real too, particularly for infrastructure roles: performance, concurrency and correctness under load. What distinguishes the loop is the insistence on rigour - a correct answer with hand-wavy justification is treated as incomplete, and being asked to prove the bound you just asserted is normal.
90 min
Hard algorithmic problems, frequently with tight constraints.
45-60 min
Two to three rounds at high difficulty with proofs of correctness and complexity.
45 min
Probability, combinatorics and expected-value reasoning.
60 min
Performance, concurrency and correctness, weighted by role.
Among the highest in the market. Problems sit at the hard end and interviewers expect you to justify correctness and complexity rather than assert them - working code with a hand-wavy argument is treated as an incomplete answer.
Not for technology roles. Probability and combinatorics are expected because the reasoning transfers, not because you need to know the products.
Practise expected-value reasoning and conditional probability until you can set the problem up cleanly. Interviewers care about the setup and the argument far more than a remembered answer to a known puzzle.
Practise the rounds D. E. Shaw actually runs, and get scored on where you stand.