Jan Deinhard

C++ Software Engineer · Embedded Linux · Real-Time & Audio · Cross-Platform Applications

C/C++ for embedded Linux, real-time audio and cross-platform applications – with more than 25 years of experience.

Discuss your project

Available from March 2027 · Remote preferred

Industries: Pro audio · Media & streaming · IoT · Industrial automation · Music software

Services

C/C++ Development

  • Modern C++, multithreading, low latency
  • New features and applications, from concept through implementation and automated testing (Catch2, Google Test) to delivery, also to reinforce an existing team
  • Modernizing existing codebases: moving grown code to modern C++ (C++11 to C++23) step by step, without putting production at risk

Embedded Linux

  • Yocto (images, layers, packages; BSPs including NXP i.MX, Raspberry Pi, BeagleBone Black, Rockchip)
  • User-space software for devices
  • Building and maintaining systems or porting them to new hardware: board bring-up, kernel configuration, custom layers, reproducible builds and CI/CD for embedded targets

Real-Time & Audio

  • Debugging real-time systems: systematically narrowing down latency, dropouts, timing problems and race conditions in concurrent C++ code (perf, Valgrind, GDB, strace) and fixing the root cause rather than working around it
  • Audio streaming, synchronization, Dante/AES67
  • Low-latency audio over IP and Wi-Fi: protocol design, buffering, clock synchronization (PTP/IEEE 1588), codec integration (Opus), ALSA, WebRTC
  • Acoustic echo cancellation and noise suppression (AEC/NS)
  • Audio plugins and virtual instruments (VST/AU/AAX) with JUCE for Windows and macOS, from real-time audio processing to a shippable plugin
  • Background: Sennheiser Mobile Connect (under 80 ms latency for 50+ concurrent users), bachelor’s thesis on clock synchronization in audio networks, US patent on media streaming, about five years of music software development at UJAM (most recently the Usynth synthesizer series)

Cross-Platform

  • A shared C++ codebase for desktop, mobile and embedded (Linux, Windows, macOS, iOS, Android), connected to native UIs (Swift/Objective-C, Kotlin/Java via JNI), as for the Mobile Connect apps at Sennheiser
  • Porting existing applications to additional operating systems, with a unified build (CMake, Conan) and CI for all target platforms
  • Connecting embedded devices, mobile apps and cloud services through clearly defined interfaces

About me

Freelance C++ software engineer with more than 25 years of experience in audio, embedded and cross-platform software. Most of my work runs close to the hardware under hard real-time constraints: low-latency audio streaming, devices running embedded Linux (Yocto) and control systems. Just as often, I write C++ code that runs on every platform: Linux, Windows, macOS, iOS and Android.

As lead engineer at Sennheiser, I helped develop Mobile Connect, a Wi-Fi-based audio streaming system that is now in use worldwide. The US patent “Interactive Media Streaming”, on which I am the first-named inventor, came out of my work at UJAM. My bachelor’s thesis was on clock synchronization for audio networks on Linux.

Most recently, I was a principal engineer at DAZN, responsible for interactive real-time features used by millions, working with teams in Israel, India, the UK and Poland. The technology there was web and cloud, but the problems were the same as in embedded systems: latency, concurrency, scale. For new projects, I am returning to C++.

Experience

  1. 06/2022–11/2026

    DAZN · Principal Engineer / Solution Architect

    Interactive real-time features for a global sports streaming platform

    Details

    Employed · Media / sports streaming · Remote

    • Architecture of interactive real-time features (FanZone) for a global sports streaming platform: chat, polls, quizzes, a moderation module and automatic profanity filtering
    • System design for live event data flows: distributing event data to millions of clients with minimal end-to-end latency
    • Technical leadership across 3–4 development teams (at times more) in Israel, India, the UK, Poland and other countries
    • Hands-on implementation in TypeScript/React in the early phase; later proofs of concept, mentoring and onboarding

    Technologies: TypeScript, JavaScript, Node.js, React, Python, AWS, REST, WebSockets, PubNub, Claude Code

  2. 09/2021–06/2022

    UJAM · Senior Software Developer

    Synthesizer plugins (Usynth), Yocto for a hardware synthesizer

    Details

    Employed · Music software · Remote

    • Usynth: a series of synthesizer plugins (VST/AU) for Windows and macOS in C++ with JUCE, plus an Electron-based installer platform for the synths
    • Embedded Linux for a hardware synthesizer with keyboard from a well-known manufacturer: Yocto configuration with a custom layer and custom image for Raspberry Pi and Rockchip

    Technologies: C++17, JUCE, VST/AU, JavaScript, Electron, Yocto, Raspberry Pi, Rockchip

  3. 12/2020–06/2022

    Texel / DAZN Israel · Software Engineer (freelance)

    WebRTC watch party with echo cancellation and noise suppression

    Details

    Sports streaming · Remote (Herzliya)

    • Texel (12/2020–10/2021): start-up for real-time communication on streaming platforms; after the acquisition by DAZN, freelance for DAZN Israel from 11/2021
    • WebRTC-based watch party system that lets users watch live sports together while talking to each other
    • Acoustic echo cancellation and noise suppression (AEC, NS): removing the TV sound picked up by the microphone during the watch party from the voice channel
    • Low-latency, frame-accurate video synchronization and peer-to-peer connectivity (iOS, Android)
    • Side engagement alongside my employment at Kieback&Peter and UJAM

    Technologies: C++17, WebRTC, Java/Android NDK, Objective-C/iOS

  4. 01/2020–08/2021

    Kieback&Peter · Lead Engineer

    Embedded Linux (Yocto) and C++ services for building automation

    Details

    Employed · Building automation · Berlin

    • Embedded systems development for building automation and industrial IoT
    • Design and implementation of Yocto-based embedded Linux platforms and C++ services
    • Setting up CI/CD pipelines and automated builds for embedded Linux targets
    • Technical lead of a small development team

    Technologies: C++17, Yocto, Node.js, CMake, GitLab CI

  5. 09/2014–12/2019

    Sennheiser Streaming Technologies · Lead Software Engineer

    Mobile Connect, low-latency audio streaming over Wi-Fi

    Details

    Employed · Pro audio · Hamburg

    • Mobile Connect: Wi-Fi-based assistive listening system for low-latency audio streaming to smartphones, now in use worldwide
    • Proprietary network protocol with clock synchronization: under 80 ms latency for 50+ concurrent users
    • Real-time audio pipeline under hard timing constraints: DSP and codec integration (Opus), ALSA, multi-stage thread architecture
    • Architecture of the embedded Linux firmware (Yocto) for custom audio hardware, from board bring-up to series production
    • Cross-platform C++ framework for the mobile apps (iOS, Android) and an admin backend in Node.js
    • Technical leadership in an international team, coordinating internal and external developers, Scrum Master, mentoring and code reviews

    Technologies: C++11/14, C, Objective-C, Java/Android NDK, Node.js, Yocto, Buildroot, ALSA, Opus, PTP/IEEE 1588, Dante, Raspberry Pi

  6. 07/2010–08/2014

    UJAM · Senior Software Developer

    Cloud DAW, audio plugins, US patent

    Details

    Employed · Music software · Bremen

    • Cloud-based, simplified DAW for hobby producers with distributed audio rendering on AWS (EC2, S3, CloudFront)
    • Architecture of a distributed audio processing system for real-time user interaction with rendering in the cloud
    • Audio plugins and virtual instruments (VST/AU) for Windows and macOS in C++ with JUCE; cross-platform C++ library for web and mobile apps
    • Audio streaming servers (C++, C#/.NET), REST services (Node.js) and iOS apps (Objective-C) for in-house products and international clients
    • Building and operating the Linux-based cloud infrastructure on AWS: architecture, deployment, automatic updates and backups
    • First-named inventor on the US patent “Interactive Media Streaming”
    • Products include UJAM Studio, VJAM, SongCruncher, Hamtaro (Propellerhead)

    Technologies: C++11, Boost C++ Libraries, JUCE, VST/AU, C#/.NET, Node.js, Python, Objective-C, AWS

  7. 05/2002–07/2010

    ViDEC Data Engineering · Software Developer

    Remote alarm notification and data transmission for industrial plants

    Details

    Employed · Industrial automation · Bremen · preceded by an apprenticeship at the same company, 1999–2002

    • Remote alarm notification software for industrial plants, developed over several product generations
    • High-performance client/server solution for transmitting plant data via GPRS; real-time communication protocols (TCP/IP) for control systems
    • Licensing and software protection system for ViDEC products
    • Responsible from requirements analysis and architecture design through to maintenance; documentation and manuals, technical customer support

    Technologies: C++, Boost C++ Libraries, C#/.NET, Windows, TCP/IP, Modbus, OPC

All positions except Texel / DAZN Israel were permanent employment.

Personal projects

Open-source projects I am currently working on. All on GitHub at github.com/polaris.

aes67-stack

C11 · MIT license · since 11/2026Work in progress

A small, portable AES67 implementation: receiving and sending AES67 audio streams on an embedded Linux board, interoperating with Dante devices in AES67 mode.

  • Core without operating system calls, dynamic memory or global state; also builds for Cortex-M microcontrollers
  • Test setup: Yocto 6.0 for the BeagleBone Black (TI AM335x) with a custom BSP layer that enables the PTP hardware timestamping unit (CPTS)
  • PTPv2 synchronization with linuxptp against Dante AVIO devices in AES67 mode: 20–50 ns RMS offset through a managed switch

Test setup: github.com/polaris/bbb-aes67-lab

vergil

C++20 · GPLv3 · since 09/2026

Low-latency, synchronized audio streaming over IP multicast on a local network: any number of receivers play the audio at the same moment, with a fixed, configurable latency.

  • PTP-style clock synchronization; adaptive resampling (windowed-sinc filter) compensates for drift between the audio clocks
  • Real-time safe: no locks or memory allocations on the audio threads, lock-free SPSC ring buffers, real-time thread priorities
  • Runs on Linux, macOS and Windows; tests with Catch2, clang-tidy, ThreadSanitizer build

remote

C++20 · MIT license

Typed RPC (msgpack-rpc) built on Boost.Asio coroutines: procedures are declared once, and client and server are checked against the same signature at compile time, with no IDL or code generator.

  • Asio-style API with completion tokens (coroutines, futures, callbacks), per-call cancellation and timeouts, pipelining
  • Robust against bad input; interoperable with msgpack-rpc implementations in other languages
  • CI on Linux (GCC, Clang), macOS (Apple Clang) and Windows (MSVC), tests under AddressSanitizer, UBSan and ThreadSanitizer

hammurabi

C++23 · MIT license · since 10/2026Work in progress

A distributed key-value store based on the Raft consensus algorithm.

  • Raft core without I/O, following Ongaro/Ousterhout: leader election, log replication, commit index; client sessions so that each write is applied at most once
  • Simulation with a simulated clock, simulated network and fault injection; automatic checking of the Raft invariants, including in chaos scenarios
  • Next: persistence, the real network with Asio, fast linearizable reads (ReadIndex)

Technologies

C++, embedded Linux, Yocto, real-time, audio, cross-platform

C++
  • C++11 to C++23
  • multithreading
  • Boost
  • Asio
  • Qt
Build & quality
  • CMake
  • Conan
  • Catch2
  • Google Test
  • Clang-Tidy
  • sanitizers (ASan, UBSan, TSan)
  • GitLab CI
  • Docker
Debugging & performance
  • GDB
  • perf
  • Valgrind
  • strace
Embedded Linux
  • Yocto
  • U-Boot
  • kernel configuration
  • board bring-up
  • cross-compilation
  • NXP i.MX
  • Raspberry Pi
  • Rockchip
  • TI AM335x
Audio & real-time
  • ALSA
  • Opus
  • Dante
  • AES67
  • PTP/IEEE 1588 (linuxptp)
  • WebRTC
  • echo cancellation and noise suppression
  • JUCE (VST/AU/AAX)
Networking
  • TCP/IP
  • UDP/multicast
  • RTP
  • WebSockets
  • ZeroMQ
  • MQTT
  • Protobuf
  • MsgPack
Platforms
  • Linux
  • Windows
  • macOS
  • iOS
  • Android (NDK/JNI)
Other languages
  • C
  • Python
  • Bash
  • TypeScript

Patent, education & certifications

Patent

US 8,745,259 B2 – “Interactive Media Streaming”
Filed August 2, 2012, granted June 3, 2014
Inventors: Jan Deinhard (first-named), Tobias Lensing, Paul Kellett · Assignee: UJAM Inc.

Live rendering system for interactive media streaming: client-side buffering, prioritized loading from the current playback position and discarding of stale data when the stream changes, for uninterrupted playback.

Patent on Google Patents

Education

  • B.Sc. in Digital Media, Wilhelm Büchner University of Applied Sciences, Darmstadt, 2013–2017, final grade 1.5 (“very good”), specialization in game development. Thesis: “Clock Synchronization for Audio Networks on Linux” (C++, ALSA, PTP, audio over IP, Yocto, NXP i.MX6UL)
  • State-certified computer scientist (Staatlich geprüfter Informatiker), continuing education at DAA-Technikum Essen, 2005–2009
  • IT specialist for application development (Fachinformatiker Anwendungsentwicklung), apprenticeship at ViDEC Data Engineering, final examination before the Bremen Chamber of Commerce, 1999–2002

Certifications

  • Dante Certification Level 3, Audinate, 11/2018
  • iSAQB® Certified Professional for Software Architecture, Foundation Level (CPSA-F), 04/2019
  • Supervised Machine Learning: Regression and Classification, DeepLearning.AI (Coursera), 01/2024
  • Advanced Learning Algorithms, DeepLearning.AI (Coursera), 02/2024

Further training

  • Modern C++1x Design and Writing Fast Code, with Andrei Alexandrescu, QA Akademie, 10/2015
  • oose seminars on software architecture, 2018–2019

How I work

  • Remote preferred; on-site meetings by arrangement (worldwide)
  • Full-time or part-time; from short engagements, such as debugging a real-time system, to long-term projects
  • I clarify requirements and constraints first. I write down architecture decisions so the team can follow them later.
  • I test automatically (Catch2, Google Test) and use CI (GitLab) and static analysis (Clang-Tidy).
  • I get up to speed quickly in existing codebases, including older ones, and modernize them step by step instead of rewriting everything.
  • If a deadline is at risk, I say so early.
  • Experienced in internationally distributed teams (Israel, India, the UK, Poland and others) with different working cultures
  • Languages: German (native), English (fluent)
  • Professional liability insurance in place
  • Open to new industries when C++, embedded Linux, real-time or cross-platform work is at the core

Contact

Available from March 2027 · Remote preferred

Discuss your project

Email
jan.deinhard@gmail.com
LinkedIn
linkedin.com/in/jandeinhard
GitHub
github.com/polaris