Steve Verhelle

Software / Systems Engineer - Go, Python, Protocol Buffers - Distributed systems

Good software engineering is not a collection of techniques - it is a disposition toward systems.

Distributed Systems Engineer with over two decades of experience architecting information systems and administering declarative environments.

I build performant, scalable, and maintainable tools by leveraging the unique strengths of a deliberate core stack - high-concurrency services in Go, data and automation in Python, contracts enforced by Protocol Buffers. Each layer earns its place.

At the systems layer, the same principles apply: explicit over implicit, reproducible over fragile, observable over opaque. I design and operate infrastructure that treats configuration as code and repeatability as a hard requirement - NixOS for declarative host configuration, Terraform for cloud provisioning, Ansible for orchestration, QUIC and SDN at the network boundary. The goal in each case is a system whose state is fully described, reproducible from scratch, and auditable.

My operating thesis: systems engineering and software engineering are not separate disciplines but two expressions of the same underlying concern. Build a system without writing software and you are assembling someone else's decisions. Write software without understanding the system it runs on and you are optimizing in the dark. In practice, serious work in either domain demands competence in both - the ratio shifts by context, but the dependency is constant.

That dual fluency shapes how I approach problems: from the machine upward and from the architecture downward, until the two meet cleanly.

Résumé

A Distributed Systems Engineer with a focus on systems programming and declarative infrastructure. Adept in applying the scientific method to system design - utilizing hypothesis-driven development and empirical observation to ensure objective, high-reliability results. Dedicated to continuous improvement and creative technical orchestration.

Skills
Language
Bash C / C++ Go HTML / CSS Javascript Nix PL/SQL Python
Platform
Ansible AWS HashiCorp Prometheus / Grafana Docker Unreal Engine Ethereum
Operational
Project management Regulatory filings Technical writing Bookkeeping Equities analysis
Aspirational
Digital Signal Processing Rust LLM training HPC administration Cryptographic algorithms French
Experience
Software / Systems Engineer
Independent & Open Source
2022 – Present
  • Designing an event-driven high-concurrency daemon supervisor system in Go.
    • Satisfies all kernel obligations when ran as the init process.
    • Python daemon configuration files.
    • Protocol Buffer communication over QUIC between the control and operational planes.
  • Developing an agentic harnesses.
  • Prototyping long-term agentic memory systems.
Equities and Currency Trading
Independent
July 2018 - June 2022
  • Development of financial calculators in Go and Python.
  • Development of MetaTrader algorithmic trading strategies.
Owner / Operator
Midnight Toll, LLC
November 2014 - July 2018
  • Responsible for all daily business operations.
  • Sole developer of a video game for mobile platforms.
  • Deployed Canonical MAAS (Metal As A Service) to provide a PXE, DNS, and DHCP environment for remote management of virtual machines.
  • Deployed Prometheus backed Grafana system monitoring dashboard.
  • Deployed a prototype HashiCorp cluster consisting of Consul, Vault, and Nomad.
  • Developed an Ansible playbook infrastructure for actualization of deployed systems.
  • Developed a Terraform module infrastructure for testing of a hybrid cloud deployment.
Caregiver
December 2010 - June 2014
  • Provided assistance with medical care and daily tasks of a family member.
Software / Systems Engineer
University of Michigan Hospital System - LMS, Ann Arbor
January 2006 - June 2006
  • Performed routine maintenance of the Learning Management System. A multi-host deployment providing a system for internal certification of hospital employees.
  • Implementation of reports generated through direct access of the Oracle database backing the Learning Management System.
Software / Systems Engineer
University of Michigan Hospital System - MCIT, Ann Arbor
December 2001 – February 2005
  • Responsible for coordination and creation of configuration files for Core Image rollouts and single workstation deployment.
  • Granting network access for non-Core devices.
  • Perform maintenance on the VTAM Web Application; an ASP, JavaScript, and VBS based front-end for an Oracle workstation configuration and IP assignment tracking database.
  • Tracking and reporting of compromised workstations and servers in the UMHS network environment.
  • Monitoring of network infrastructure.
  • Co-development of the Core Workstation Database Web Application, a Coldfusion / Javascript replacement for the VTAM Web Application.
Logistics Clerk
ACC Penske, Flint
March 2001 - June 2001
  • Responsible for updating and maintaining data for GM on-demand warehouse inventory system.
  • Logistical tracking of part containers for warehouse.
  • Processing and approval of shipping and receiving paperwork.
Mechanic
Prestige Chevrolet, Flushing
March 2000 – September 2000
  • Performed routine maintenance and inspection of vehicles.
  • Performed diagnosis and repair of mechanical and electrical failures.

Core Knowledge

unix

Advanced system administration

From at to zcat. From DNS to email. I possess the skills to configure and administer your systems.

clusters

Distributed systems architecture

Whether it is consensus or discovery I can navigate a route through the hive.

programming

Protocol Buffers

Protobufs maintain strictly typed, language-agnostic, and lightning-fast communication contracts.

programming

Advanced software design

From build tooling to design patterns I can craft anything given the resources.

unix

NixOS

A unique, declarative Linux distribution built on top of the Nix package manager, where the entire operating system configuration - including packages, system services, and configuration files - is managed via a single functional language.

systems

Advanced networking

Software-Defined Networking (SDN), Cloud and Data Center Networking, Websockets, and QUIC are of prime interest.

programming

Go

Powering high-concurrency backend services, CLI tools, and systems that require raw performance and fast execution.

systems

DevOps

Whether it is NixOS or Terraform and Ansible I have the experience to overcome your deployment gaps.

programming

Python

The province of data processing, scripting, automation, and rapid prototyping within the ecosystem.

emerging

AI

Applying declarative system architectures to constrain, orchestrate, and bring reproducibility to probabilistic AI models.

Go and Python with Protocol Buffers is a proven stack for high-performance, cross-language microservices. Go owns performance-critical services; Python handles data science, ML, and automation. Communication is anchored by centralized .proto files defining shared data structures and service contracts, from which the toolchain automatically produces the code each language needs to send and receive structured messages.

The breadth above is not accidental. Sustained curiosity - the habit of pulling threads past the point of immediate utility - is what connects distributed systems to NixOS to AI agent orchestration. Exploration and experimentation are the mechanism; the knowledge is the residue.

Projects

Epilogue

The best systems are the ones nobody notices - until they're gone.

Good software engineering is not a collection of techniques - it is a disposition toward systems. It is the habit of asking what a thing actually does before asking how to make it faster, the reflex to name failure modes before celebrating success paths, and the discipline to leave a codebase more legible than you found it. These are not taught in a curriculum. They are earned through years of maintaining systems that outlive their original authors, debugging production failures at 2 AM, and learning - repeatedly - that complexity is not a sign of sophistication but of deferred decisions.

I have spent more than two decades cultivating that disposition. My career traces the full vertical of modern infrastructure: from the machine-level concerns of configuration management and network provisioning, through the architectural challenges of distributed systems and high-concurrency services, to the emerging frontier of agentic AI. What connects those layers is not the technology but the underlying commitment to systems that are correct, observable, and honest about their failure modes.

Information technologies are equally prominent throughout my personal endeavors. Whether it is the daisy-chained topographies of MIDI (Musical Instrument Digital Interface) networks. Or gain staging - the process of managing audio signal levels at each stage of a recording or mixing chain to ensure optimal sound quality, maximize headroom, and minimize noise and distortion. Music production and engineering, both theory and practice, are just systems for me to study, engage, and re-engineer.

The same instincts that drive my engineering practice show up in how I make music. I prefer hardware synthesis over software production not out of nostalgia but for the same reason I prefer explicit systems: tactile hardware forces you to commit. A patch cable is a contract. A knob position is a decision. There is no undo history on a modular synthesizer, and that type of constraint produces a different kind of attention. The machine pushes back, and the push-back is the point.

If you have read this far, thank you. The time you spent here is not something I take lightly. Whether you are evaluating a hire, exploring a collaboration, or simply curious - I hope the document gave you an honest picture.