Principal Low-Latency Systems Engineer (C++ / HFT)

ADVANCE SOFTWARE ENGINEERING PTE. LTD.; Singapore; On-site; SGD 250,000 - 420,000; Full time

Job summary

ADVANCE SOFTWARE ENGINEERING PTE. LTD. in Singapore seeks a Principal Systems Engineer to architect ultra-low-latency trading infrastructure.

You will push the hardware-software boundary and optimize critical paths with C++ and Linux internals. You will design lock-free data structures, kernel-bypass networking, and high-performance profiling to zero in microsecond bottlenecks on nanosecond timescales. This is a demanding role for elite engineers with a track record in HFT, extensive C++

Qualifications

• 8+ years of heavyweight C++ (C++17/20) experience with deep STL knowledge.
• Proven track record building sub-microsecond tick-to-trade platforms in finance.
• Expertise in Linux kernel internals, NUMA, and CPU pinning; familiarity with FIX/OUCH/ITCH protocols.

Responsibilities

• Architect ultra-low-latency trading engines, pricing models, and order execution gateways.
• Optimize hardware-software path with deep memory access and cache optimization.
• Engineer kernel-bypass networking and low-latency network protocols (TCP/UDP/Multicast).
• Design lock-free, wait-free data structures and bespoke allocators for deterministic timing.
• Use perf/VTune/Valgrind to identify and remove microsecond bottlenecks in production.

Skills

C++ mastery

HFT domain expertise

Linux kernel knowledge

Network programming

Lock-free data structures

Performance profiling

Education

Master's degree in CS/Math/Physics

Tools

Solarflare OpenOnload

DPDK

RDMA

Job description

The Objective

We are seeking an elite, uncompromising Principal Systems Engineer to architect and optimize the critical path of our ultra-low-latency algorithmic trading infrastructure. This role commands a premium compensation package and demands unparalleled technical rigor. This is not a standard software engineering role; we operate in nanoseconds. You must possess absolute mastery of modern C++, hardware sympathy, and Linux OS internals. There is zero margin for garbage collection pauses, unoptimized cache lines, or inefficient context switches. You will write code where every CPU cycle is accounted for.

Core Responsibilities

• Nanosecond Engineering: Architect, build, and relentlessly optimize ultra-low-latency trading engines, pricing models, and order execution gateways.
• Hardware Sympathy: Dictate system architecture with a profound understanding of underlying hardware architectures, mastering L1/L2/L3 cache utilization, branch prediction, and memory access patterns.
• Network & Kernel Optimization: Eradicate latency by engineering kernel-bypass networking solutions (Solarflare OpenOnload, DPDK, RDMA). You must dominate raw socket programming and packet-level network protocols (TCP, UDP, IP Multicast).
• Lock-Free Concurrency: Design infallible, thread-safe, lock-free, and wait-free data structures. Implement bespoke memory allocators to guarantee deterministic execution times along the critical path.
• Performance Profiling: Command advanced profiling tools (perf, VTune, Valgrind) to identify and destroy microsecond-level bottlenecks in production environments.

Strict Qualifying Criteria

• Unrivaled C++ Mastery: Minimum of 8+ years of heavyweight, battle-tested experience in modern C++ (C++17/20). You must know the standard library deeply, and more importantly, know exactly when not to use it.
• HFT Domain Expertise: Proven track record building profitable, sub-microsecond tick-to-trade platforms in a proprietary trading firm, hedge fund, or top-tier financial institution.
• Linux OS Dominance: Expert-level knowledge of Linux kernel internals, CPU pinning, NUMA architecture, and thread affinity.
• Protocol Authority: Absolute proficiency in FIX, OUCH, ITCH, and binary exchange protocols.
• Academic Pedigree: Preferably a Master's in Computer Science, Mathematics, Physics, or a strictly equivalent quantitative discipline from a top-tier university is aggressively preferred.