>
L
Cat
June 7, 2026 - Ghent, Belgium
Openv26.06

Liam Goethals

An AI engineer optimized for product-facing GenAI systems: LLM orchestration, agentic systems, evaluation, trace analysis, prompt iterations. Available under a conversation-first license.

liam-goethals-4+1
AI Engineering
Product building
Stack
PythonTypeScript
Deploymenti
on-site-compatible
Spokeni
Dutch

/fluent

English

/fluent

French

/very-basic

Features

LLM orchestration
Structured outputs
Prompt systems
Eval loops
Trace analysis
Efficient context usage
Product experiments
Retrieval workflows
Backend APIs
AI strategy
Applied research
Frontier pretraining
Unmeasured shipping

Greyed capabilities are intentionally unsupported. See known limitations.

Training data

Mechanics & metalworking

Questionable entrypoint into AI? Perhaps. Relevant nonetheless.

Education

Computer science: Bachelor in Applied Information Technology - specialised in Artificial Intelligence & data engineering (HOGENT).

Backend systems

APIs, data flows, reliability, and the usual software and scaling challenges.

Photography & Videography

Analog, digital, moving, stills, commercial, creative.

In The Pocket

AI strategy and GenAI product building across messy real-world contexts: Flemish and federal government, healthcare, ports and logistics, retail, and scientific publishing.

Springer Nature

Helped shape and build Springer Nature’s largest GenAI product initiative to date: a long-running applied AI engagement with outstanding researchers.

Donna

Production GenAI in a scaling startup: international rollout, 20+ languages, quality, scale, fast iterations, recursive self-improvement, close user connection.

Use cases

Recommended use cases

Recommended for

AI product teams where model behavior has to become product quality.

  • chat
  • documents
  • search
  • agents
  • workflow automation
  • recursive self-improvement
  • evals
  • observability
Not recommended for
  • Pure frontier pretraining.
  • Demo's without futures.
  • Unhinged vibecoding.
Limitations

Known limitations

  • 01

    Most useful where research meets product, less so at the pure-frontier extreme.

  • 02

    Has done real research, but is usually happiest when it ends up shipping.

  • 03

    Allergic to avocados.

History

Deployment history

Donna

AI Engineer

2025 to now

Production GenAI product, 20+ languages, growing team, real users.

In The Pocket

AI Engineer

2022 to 2025

AI strategy and product experimentation across client contexts, including government and public-sector programs.

WeGroup

Backend Developer

2019 to 2022 · Student role

Backend foundations while studying computer science.

Other versions

Backend Liam

v23.0

Reliability, APIs, and data flows.

Photographer Liam

v-

35mm.

Metalworker Liam

v16.0

Welding and stuff.