Akuna's coding challenges are often considered slightly less math-heavy than Jane Street but more focused on low-latency and systems thinking than generic FAANG. Expect 2-3 medium to hard LeetCode-style problems with a twist: they frequently involve data structures for efficient processing (like heaps, hash maps) and may include follow-ups on optimizing for time/space complexity in a trading context. The key differentiator is the emphasis on clean, production-quality code and discussing trade-offs.
Master arrays/strings, hash maps/dictionaries, sliding window, two pointers, binary search, trees (especially Tries for prefix problems), heaps, and graphs (DFS/BFS). Pay special attention to problems involving real-time data streams or order book simulation, as these mirror trading scenarios. Practice explaining your choice of data structure for a given problem's constraints—they value this analytical discussion.
The Bar Raiser is a behavioral and leadership圆interviews round, similar to Amazon's model, where a senior interviewer evaluates your alignment with Akuna's core principles (like 'Customer Obsession' and 'Ownership'). Prepare 5-6 detailed STAR stories that demonstrate these principles, especially around collaboration, handling ambiguity, and making data-driven decisions. Research Akuna's specific values on their careers page and tailor your examples to show how you've lived them in past projects or internships.
The timeline can vary significantly. Typically, you might hear back within 2-4 weeks, but it's not uncommon for the process to take longer, especially during peak recruiting seasons. If you haven't heard after 3 weeks, a polite follow-up email to your recruiter is acceptable. Delays often occur due to committee reviews and team-matching, not necessarily a negative signal.
SDE-1 focuses heavily on core DSA and clean implementation. SDE-2 adds a dedicated system design round (e.g., design a low-latency market data feed) and expects more depth in one technical domain. SDE-3 involves architectural system design (trade-offs, scalability, fault tolerance) and often a deep-dive into your past project leadership. All levels include the Bar Raiser, but senior roles expect more demonstrated impact and mentorship examples.
The biggest mistake is writing code without first fully clarifying requirements and edge cases. Candidates also often fail to discuss time/space complexity continuously or consider real-world constraints like memory limits. Another common error is neglecting to test the code with examples before finishing. Always verbalize your thought process, write modular code, and proactively discuss how you'd extend the solution.
For SDE-1/2, prioritize LeetCode (150-200 problems, mixing mediums and hards) as the gatekeeper. For SDE-2/3, supplement with systems concepts: understand TCP vs. UDP, kernel bypass (e.g., DPDK), cache coherence, concurrent data structures, and garbage collection pauses. Read Akuna's engineering blog for their tech stack (C++, Python, FPGA) and be ready to discuss how you'd optimize a simple problem for nanosecond latency.
Go beyond generic 'I love finance.' Mention specific aspects like their two-sided market making model, risk management challenges, or how their technology stack (e.g., use of FPGA for risk checks) aligns with your skills. Ask insightful questions about technical challenges in their stack, the deployment pipeline for trading strategies, or how they measure latency percentiles. This shows you've researched their business model and aren't just applying to any trading firm.
