YAML Metadata Warning:empty or missing yaml metadata in repo card

Check out the documentation for more information.


Jonas Falk

Location: Gothenburg, Sweden (UTC+1 / UTC+2)
Focus:    Low-Latency Systems & Digital Signal Processing (DSP)
Stack:    Rust, Modern C++ (C++20), Zig, Linux Real-Time Audio
Status:   Building open audio spatialization tooling & remote acoustic telemetry

About Me

I am a senior systems software engineer specializing in deterministic real-time audio architectures, spatial acoustics, and high-throughput hardware optimization. My engineering practice centers on crafting zero-allocation audio processing pipelines, designing lock-free concurrent primitives, and structuring memory layouts for optimal SIMD vectorization.

When not profiling ring buffers with hardware counters or measuring filter phase linearity with an oscilloscope, I spend my time field-recording avian acoustic patterns across the Swedish west coast or restoring vintage analog reel-to-reel audio electronics.

  • 🔭 Designing deterministic, lock-free real-time audio graphs.
  • ⚡ Optimizing SIMD kernels (AVX-512 & ARM NEON) for partitioned impulse convolution.
  • 🛡️ Hardening embedded acoustic sensor firmware with strict memory isolation.
  • ☕ Powered by single-origin light roasts and reproducible build systems.

Technical Stack & Tooling

Domain Primary Technologies Practical Utilities
Systems & Real-Time Rust, C++20, Zig, C (C11) eBPF, Valgrind, perf, criterion.rs
Audio DSP & Spatialization HRTF Filters, B-Format Ambisonics, SIMD Convolver JACK Audio, ALSA, PortAudio, WebAudio
Embedded & Hardware STM32 (Cortex-M4/M7), ESP32-S3, KiCad I2S Codecs, Logic Analyzers, Oscilloscopes
Environment & Infrastructure Arch Linux, NixOS, Alpine Neovim, CMake, Ninja, Git, Docker

Featured Open Source Projects

solstice-dsp

Zero-allocation, header-first spatial audio convolution library written in Rust.

  • Core Architecture: B-format Ambisonics decoder featuring SIMD-partitioned FFT convolution.
  • Latency Guarantee: Sub-1.2ms processing budget at 96kHz / 64 buffer samples without dynamic heap allocation.
  • Targets: Native desktop (Linux/macOS/Windows) and WebAssembly target for interactive synthesizers.

pulsefield

Ultra-low power acoustic telemetry daemon for remote ecological listening stations.

  • Hardware Integration: Custom STM32 firmware communicating via LTE-M / LoRaWAN with fall-back SD card ring buffer.
  • Compression: Adaptive FLAC frame packing optimized for narrow-bandwidth intermittent solar bursts.
  • Resilience: Hardware watchdog timers and fault-tolerant state recovery routines designed for sub-zero Nordic winters.

decibel-tui

Terminal-based broadcast loudness analyzer adhering strictly to ITU-R BS.1770-4 and EBU R128 standards.

  • User Interface: Keyboard-driven ratatui interface with real-time true-peak metering and LUFS integration logs.
  • Driver Support: Direct ALSA and CoreAudio backend hooks without intermediary desktop mixing layers.

Engineering Principles

  1. Deterministic Hot Paths: The audio processing callback must never allocate memory, acquire locks, or perform non-deterministic disk/network I/O.
  2. First-Principles Understanding: Always understand the hardware cache hierarchies, instruction latency, and transistor limits before introducing high-level abstractions.
  3. Respect for Physical Longevity: Write modular, dependency-minimal code that can compile and operate unmaintained on low-power silicon a decade from now.

Contact & Public Keys

  • Website / Tech Notes: https://jonasfalk.dev
  • Email: jonas [at] jonasfalk.dev
  • Matrix: @jonas:matrix.org
  • GPG Fingerprint: 7E9B 4410 82C9 52A1 D84F 2F0A 1B83 99C0 411E 32A5
Downloads last month

-

Downloads are not tracked for this model. How to track
Inference Providers NEW
This model isn't deployed by any Inference Provider. 🙋 Ask for provider support