HFT / Quant SWE Roadmap — 2025 → 2027

WBD →
Optiver.
24 months.

A complete, phase-by-phase plan to transition from Software Engineer II at Warner Bros. Discovery into a mid-level quant or HFT software engineering role. Built around C++ mastery, systems depth, low-latency thinking, and a portfolio that speaks for itself.

24Months total
4Phases
400+DSA problems
€250kCeiling (quant, strong year)
85%Hire probability at 24 months
Phase 1 Foundations Months 1 – 6 C++ fluency, DSA fundamentals, Linux & OS basics, first concurrency patterns. Building the base everything else rests on. Interview probability: 10 – 15% Phase 2 Systems & Performance Months 7 – 12 Advanced C++, profiling tools, OS internals, networking, trading domain knowledge, and your first portfolio project. Interview probability: 40 – 50% Phase 3 Low-Latency Mastery Months 13 – 18 Lock-free programming, hardware-level optimisation, kernel bypass, HFT architecture, full portfolio expansion. Interview probability: 60 – 70% Phase 4 Polish & Acquire Months 19 – 24 Consolidate all knowledge, finalise portfolio, open-source contributions, networking, and targeted application strategy. Interview probability: 85 – 95%
01 Learning tracks — run across all phases
C++ Mastery From syntax familiarity to expert-level: memory model, templates, lock-free structures, hardware intrinsics. The primary language of every HFT firm.
DSA 150 problems by month 6, 300 by month 12, 400+ by month 18. C++ only. Timed solves from month 7 onwards. DP is the inflection point.
Linux & OS Processes, virtual memory, scheduling, interrupts, kernel internals. HFT systems run on bare Linux and you need to understand what happens beneath your code.
Networking TCP/IP, UDP multicast, socket programming, kernel bypass (DPDK, io_uring). Market data travels on the wire and every microsecond on that path matters.
Math & Probability Expectation, variance, distributions, Markov chains, brainteasers. Optiver and IMC filter on probability puzzles at the SWE interview level.
Trading Domain Order books, matching engines, market microstructure, HFT system architecture. You don't need to be a quant researcher — but you need to speak the language.
Portfolio Phase 1: data structures. Phase 2: limit order book. Phase 3: full matching engine + market data handler. Phase 4: documented, benchmarked, blog-posted.
WBD (day job) Find a latency problem in production. Own the fix. Instrument it. "Reduced p99 from Xms to Yms" is the single most powerful sentence on your quant CV.
02 The rule

Build things that are hard to fake. A profiled order book. A working lock-free queue. A measured latency improvement in production. These are not things you can vibe-code into existence — and that is exactly the point.

— Core principle, April 2026