fsafdlsaf/gmhmjmkmlm

KR 한국어	US English	JP 日本語	CN 简体中文	ES Español
FR Français	DE Deutsch	RU Русский	SA العربية	HI हिन्दी

---

Hey there, I'm Simon 👋

I build low-latency systems, distributed storage engines, and network primitives. Most of my work sits around the intersection of the Linux kernel, lock-free concurrency, and mechanical sympathy. When I am not optimizing memory layouts or tracing packet drops with eBPF, I tinker with analog audio synthesizers and mechanical keyboards.

• 🔭 Current Focus: Designing a zero-copy asynchronous replication layer for append-only distributed event stores.
• ⚡ Engineering Principle: Predictable p99 latency beats peak throughput. Always measure with hardware performance counters before guessing.
• 💬 Ask me about: Rust, Zig, Linux systems programming, LSM-trees, DPDK, eBPF telemetry, and audio DSP.
• 📍 Location: Stockholm, Sweden (UTC+1 / UTC+2)
• 📫 Reach out: simon.vance.dev@gmail.com

---

🛠️ Engineering Disciplines & Technical Stack

I care deeply about software that treats system memory, CPU cache lines, and disk pages with respect. Here is the toolkit I rely on daily:

Domain	Primary Tools & Technologies	Focus & Practices
Systems & Core	Rust, Zig, Modern C (C11/C23), Go	Lock-free data structures, zero-allocation paths, SIMD vectorization
Networking & Kernel	eBPF / XDP, io_uring, epoll, QUIC	Raw socket parsing, kernel bypass, custom packet pipelines
Storage & Data	RocksDB, custom LSM engines, SQLite, DuckDB	Write-ahead logging (WAL), cache-oblivious B-trees, zero-copy serialization
Infra & Observability	Linux (Arch / Debian), Docker, perf, bpftrace, Prometheus	Memory profiling (valgrind, dhat), flamegraphs, hardware PMU counters
Tooling & Shell	Neovim, WezTerm, Git, Bash/Zsh, Make, CMake	Keyboard-driven workflow, terminal-centric debugging, reproducible toolchains

---

🚀 Highlighted Open Source Projects

Here are a few infrastructure projects and experiments I've designed and maintain:

1. vortex-kv

High-performance embedded key-value storage engine engineered in Rust.

• Architecture: Append-only log-structured merge tree (LSM-Tree) backed by io_uring direct I/O.
• Highlights: Zero heap allocations on read hits, fractional cascading Bloom filters, and concurrent compaction pipelines.
• Key Metrics: Sustains 420,000+ random writes/sec on NVMe storage with deterministic sub-millisecond p99 latency.

2. meshpipe

Modular UDP/QUIC packet multiplexer and telemetry probe written in Zig.

• Architecture: Event-driven worker pool utilizing arena allocators and cacheline-aligned ring buffers.
• Highlights: In-kernel packet filtering via eBPF XDP hook with fallback to user-space SIMD checksumming.
• Use Case: Real-time distributed state synchronization across edge cluster nodes over lossy networks.

3. chronos-probe

Minimalist kernel telemetry daemon written in Go and C (eBPF).

• Architecture: Attaches tracepoints and kprobes to disk block layer and TCP stack without altering running binaries.
• Highlights: Emits zero-overhead Prometheus histograms for page fault latency and scheduler runqueue delays.
• Use Case: Diagnosing silent tail latency spikes in latency-critical production microservices.

4. poly-dsp

Lightweight polyphonic wavetable sound synthesizer engine in modern C.

• Architecture: Lock-free audio buffer ring with SSE/AVX2 vectorized audio oscillators and bandlimited BLEP synthesis.
• Highlights: Runs at 96kHz / 32-bit float with less than 1.2% single-core CPU utilization on x86_64.

---

📊 System Architecture Mindset

 [ Client Layer ]
        │  (QUIC / Custom Binary Protocol)
        ▼
 [ eBPF XDP Filter ] ─── (Early Drop / Fast Path)
        │
        ▼ (Kernel io_uring Completion Queue)
 [ Ring Buffer / SPSC Queues ] ── Cache-line Padded
        │
   ┌────┴──────────────────────────┐
   ▼                               ▼
[ Memory Arena / Off-Heap ]   [ WAL Disk Persistence ]
   │ (SIMD Index Scanning)         │ (Direct I/O O_DIRECT)
   └───────────────┬───────────────┘
                   ▼
         [ Stable State Query ]

• Mechanical Sympathy: Structure-of-Arrays (SoA) layout over Array-of-Structures (AoS) to maximize L1/L2 data cache residency.
• Allocation Discipline: Memory arenas and pre-allocated ring buffers in hot loops; zero garbage collector latency spikes.
• Failure Transparency: Explicit error handling, deterministic crash recovery via write-ahead logging, and exhaustive property testing.

---

💻 Personal Rig & Workspace Setup

• Operating System: Arch Linux (Kernel linux-zen) with Sway Wayland compositor
• Editor: Neovim (custom Lua configuration, Treesitter, native LSP)
• Terminal: WezTerm with Tmux session management
• Keyboard: Corne Split Ergonomic Keyboard (Colemak-DH layout, 42 keys, custom QMK firmware)
• Audio Gear: Arturia KeyStep Pro, Eurorack modular rig, Beyerdynamic DT 1990 Pro

---

🤝 Let's Collaborate & Connect

I am always interested in discussing storage engines, kernel telemetry, zero-copy networks, and open-source systems architecture.

Email:       simon.vance.dev@gmail.com
Matrix:      @simonvance:matrix.org
PGP Key:     8F4B 2C91 E0D3 7A88 41F2  990C 5E3A 1204 D91A B7C2
GitHub:      https://github.com/gmhmjmkmlm-rgb
Discussions: Open an issue or drop a line in GitHub Discussions

---

“Simplicity is prerequisite for reliability.” — Edsger W. Dijkstra