Thomas BrochardDevelopment · Data · AI
Let’s talk

Portfolio — Thomas Brochard

Design & development

From raw data
to useful systems.

I connect data engineering, development and AI to make complex systems clearer and easier to use.

Data engineeringApplications & interfacesRetrieval systems & AI

01 /Selected projects

Design, put into practice.

Concrete challenges.
Considered technical decisions.

Further explorations

A collection of 10 case studies
8 case studiesContext / decisions / scope

These are technical demonstrations, selected-source showcases and architecture studies. The case studies describe their scope and limitations. GitHub repositories are private and require authorized access.

02 /Method

Complexity deserves
a clear structure.

A project is understood through its decisions as much as its interface. Each of those decisions deserves to be explicit.

01

Define the right boundaries.

Clarify the need, available data and scope before choosing the tools.

02

Build reliable connections.

Define the models, contracts and checks that connect the components of a system.

03

Make decisions readable.

Document the reasoning, expose the limits and design understandable interfaces.

03 /About

The bigger picture.
An eye for detail.

I’m interested in what connects data to its uses: its structure, how it moves and the way we interact with it.

This portfolio brings together my work on data pipelines, analytics, retrieval systems and interfaces. Each study makes the technical reasoning and the project’s limits visible.

My GitHub profile

04 / Areas of practice

From the interface
to the data flow.

My work connects what a person sees with the models, transformations and workflows that support it.

01

Interfaces with clear state.

Structure, responsive layouts and interaction design. In complex interfaces, that also means preserving drafts, explaining unknown states and making concurrent edits visible.

  • HTML & CSS
  • JavaScript
  • TypeScript
  • State handling

In practice: Fly Drone Tool ↗

02

Workflows with explicit steps.

Python scripts, orchestration and batch processes. Make dependencies inspectable, keep progress visible and place quality checks where they can explain a failure.

  • Python
  • Kestra
  • Batch imports
  • Quality gates

In practice: BottleNeck ↗

03

Data with a traceable path.

Relational models, SQL transformations and analytical layers. Define the grain, preserve provenance and reconcile records before drawing a conclusion.

  • SQL
  • DuckDB
  • PostgreSQL
  • dbt
  • pandas

In practice: Weather ELT Pipeline ↗

The next conversation

Let’s talk about your project.

↗
A need, an idea, a technical question.
LinkedIn ↗GitHub ↗