Skip to content
AD

ENGINEERING SERVICES

When the failure is unclear, start with evidence.

I help teams investigate difficult Linux and native-software problems, improve failure visibility and deliver targeted software changes. Engagements can start small, remotely, with an explicit scope and useful handover.

Describe your problem

Ways to work together

A defined outcome, not a vague list of skills.

01Focused entry point

Failure Triage Sprint

For crashes, freezes, watchdog events and intermittent Linux or embedded failures.

Start with a one-day evidence review or agree on a short multi-day investigation when the symptoms warrant it.

What you receive

  • Known facts and gaps in the evidence
  • Ranked hypotheses and reproduction plan
  • Recommended instrumentation and next actions
Discuss this engagement
02Systems engineering

Debugging & Reliability Engineering

For issues requiring deeper analysis of native code, processes, I/O, memory or degraded system behavior.

Scoped work around a reproducible failure, investigation milestone or implementation of better diagnostics.

What you receive

  • Technical findings with supporting evidence
  • Targeted fixes or observability changes where feasible
  • Validation notes and remaining risks
Discuss this engagement
03Software delivery

Backend, Integration & Tooling

For teams needing an API, an integration, automation, an internal tool or a difficult existing-codebase fix.

Small, concrete deliverables with an agreed definition of done, rather than an open-ended rewrite.

What you receive

  • Implemented change or working integration
  • Tests or reproducible validation
  • Handover notes for the existing team
Discuss this engagement

How I work

A practical process with clear handoffs.

01

Share the context

Describe the system, symptoms, timeline, environment and what has already been tried.

02

Agree on scope

We define access, constraints, expected evidence, deliverables and a realistic timebox.

03

Investigate or build

The work follows observable facts, short feedback loops and explicit checkpoints.

04

Hand over results

You receive what was learned or shipped, validation notes and the next recommended steps.

Investigation cannot guarantee that every intermittent failure will be reproduced or fixed in a fixed number of days. The agreed output always distinguishes verified findings, open questions and next actions.

Evidence over claims

See the engineering work.

Review public contributions and reproducible prototypes before deciding whether the approach fits your problem.

Selected work

Technical education

Need your team trained instead?

I also deliver practical engineering workshops in French and English, tailored to the participants and their tools.

Teaching & training

Tell me what you are seeing.

A useful first message includes the runtime environment, the symptom, how often it occurs, relevant logs or traces if shareable, and the outcome you need. Please avoid sending confidential data until a suitable agreement is in place.

Email [email protected]