Legal technology
OpenLegi
OpenLegi connects legal sources to AI assistants. We help develop the product and the infrastructure it runs on.
We build software and automate business and industrial processes. From office tasks to production lines, we turn your needs into practical solutions.
€2,400/ monthExcluding VATNo commitment
Tools and automation for products and everyday operations
OpenLegiSanchez HuitresMonthly subscription
Develop your software and automate business or industrial processes as your needs change.
See how we support a legal AI product and automate a daily counting and reporting process.
Legal technology
OpenLegi connects legal sources to AI assistants. We help develop the product and the infrastructure it runs on.
Oyster farming
The team needs daily oyster counts. We develop software that uses camera images to count oysters and produces daily reports. We also provide practical AI training.
We use our software factory for software and automation. It combines AI-assisted development and integration with automated tests and human oversight.
We set priorities with you. We look at the current process, agree on the result and decide how to check it. We split large requests into smaller changes.
We build the software or configure the automation. Frédéric is accountable for the work. We review and test each change before release.
We put the change into use. If you subscribe, we meet with you for 30 minutes each week and send a monthly report of what we delivered.
Choose a subscription when your needs change over time or a fixed-scope project for an agreed result. You can start a new software or automation project either way.
Ask for a quoteWe agree on the scope and quote the work. For example, we can build a micro-SaaS product or a website with a CRM, connect systems, or automate a reporting process.
We teach your team to use Claude for the tasks they do at work. Training is quoted separately.
Previous work on software products, cloud infrastructure and tools that automate business tasks.
05/2026 → 06/2026Freelance
Emmeti Fonderia · Software Engineer
Cost optimization internal tool for an Italian bronze foundry: 9-phase PL/pgSQL cost engine, per-article routing, configurable rate tables, pgTAP test coverage.

12/2025 → 05/2026Freelance
Optimia · Chief Technology Officer
SaaS platform built from scratch as CTO, covering architecture, infrastructure, CI/CD, and team mentoring from day one
10/2025 → nowFreelance
Kigs · Senior Software Engineer, DevOps
Harness engineering for a mobile app startup: CI/CD, TDD adoption, release process reliability, plus the agent rules and skills that keep AI-assisted delivery consistent across the team.

07/2025 → nowStudio
FREDFERRE · Founder & Studio Lead
Live production AI platform for French legal contract analysis: surfaces risks, key clauses, and obligations in minutes, grounded in French employment law.
04/2025 → 07/2025Studio
FREDFERRE · Founder & Studio Lead
Open-source Python SDK for Légifrance, France's official legal document database. Almost no one publishes one.
09/2024 → 02/2025Employee
Ambrosys · Backend DevOps Engineer
Mobile toll payment app backed by a serverless API with full observability and progressive feature rollout
Private / Internal

05/2024 → 08/2024Employee
Ambrosys · Backend DevOps Engineer
Automated testing pipeline for ML models on virtual hardware with Docker orchestration and serverless infrastructure

09/2023 → 04/2024Employee
Ambrosys · Backend DevOps Engineer
Data platform for GGI-Spermex's cattle breeding program, helping dairy farmers optimize mating decisions through statistical analysis

03/2023 → 08/2023Employee
Ambrosys · Backend DevOps Engineer
Cloud infrastructure for an open-source AI workflow platform enabling ML model training and deployment on HPC clusters

04/2022 → 02/2023Employee
Ambrosys · Backend DevOps Engineer
GPS quality control platform for the German truck tolling system, automating statistical evaluation of positioning data across millions of vehicles
Feedback from clients and teams we have worked with.
Frédéric quickly understood our cost-calculation process and delivered a reliable internal tool in a few weeks. The cost engine is accurate and well tested. Serious work, exactly what we needed.
We were looking for an expert capable of developing a SaaS MVP on a high-performance stack. Frédéric delivered exactly what was needed, mastering the Scaleway / Supabase ecosystem end-to-end. The GitLab deployment pipeline configuration is exemplary and meets the highest current DevOps standards. A worthwhile investment for anyone wanting a serious platform from day one.
Mr. Ferrera performed well across all areas of his responsibilities. He demonstrated solid analytical and creative thinking skills, quickly grasping new tasks and adapting effectively to changing circumstances. His resilience under pressure, flexibility, and problem-solving capabilities were significant assets to our team.

Founder and technical lead
Frédéric Ferrera founded FREDFERRE. He is accountable for technical decisions and the software and automation we deliver. We work remotely from Montpellier.
Proficient across the entire technology stack, from responsive frontends to scalable backend systems and database design.
AWS-certified professional with expertise in cloud-native infrastructure, containerization, and automated deployment pipelines.
Harness engineering and applied AI integration: building the scaffolding (agent loops, tools, memory, permissions) that turns AI models into reliable production agents, with strong background in agentic workflows and intelligent automation.
Strong background in data pipeline architecture, real-time processing, and building analytics platforms for business intelligence.
Articles on software engineering, architecture, development practices and applied AI. We publish them on Substack.
AI EngineeringSep 20, 20266 min read
When coding agents do most of the work, a ticket splits into human coordination and agent execution, and the agents' bill is tokens, not days. Estimating a ticket becomes two questions, how much coordination and how long a run, and a feature's measured cost adds up into two lines: people and machine.
Read on SubstackClaude CodeJun 29, 20267 min read
Why a team's AI assistant configuration should be version-controlled, peer-reviewed, and shared like any other code. Once skills ship through code review and a package mechanism, a Product Owner can contribute to the toolkit every developer uses, and private setups stop drifting apart.
Read on SubstackAI EngineeringJun 8, 20269 min read
Automating the build pipeline between refined requirements and merged code: squads of Developer, QA, and Code Reviewer agents work in parallel on individual issues, with tests derived from Gherkin acceptance criteria and no planning checkpoint. Judgment stays upstream in refinement. The real lever is harness engineering, where the scaffolding around agents matters more than the model itself.
Read on SubstackAI EngineeringMay 28, 20269 min read
How AI agents reshape product refinement: human PMs collaborate with specialist agents (architect, developer, designer, analyst, QA) in real-time sessions, guarded by an ambiguity gate that blocks unresolved questions and a complexity score that forces big changes into multi-issue projects.
Read on SubstackAI EngineeringMay 5, 202610 min read
AI coding assistants are not bottlenecked by generation speed but by missing context. Inverting the workflow to spend 80% on specs and architecture turns documentation into machine-readable infrastructure for faster, more consistent delivery.
Read on SubstackClaude CodeApr 30, 20267 min read
Why ad-hoc dotfile sharing fails for AI tooling, and how a Claude Code plugin marketplace turns scattered configurations into a single source of truth the team can version, review, and ship.
Read articleClaude CodeMar 7, 202610 min read
A decision framework for structuring Claude Code's 5 context layers: CLAUDE.md, Rules, Skills, Agents, and Hooks. Interactive guide with decision tree, comparison matrix, and scope hierarchy.
Read articleAI EngineeringFeb 14, 20268 min read
How language models with structured output can turn plain-English descriptions into functional data dashboards, making visualizations disposable and on-demand.
Read on SubstackDevOpsDec 15, 202412 min read
A comprehensive guide to deploying Strapi CMS on AWS infrastructure. Learn how to set up a scalable, production-ready Strapi deployment using AWS services like EC2, RDS, and S3.
Read on SubstackYes. It covers software development and automation of business and industrial processes. Both use the same request queue.
Yes. You can start a new software or automation project on the subscription with no setup fee.
We assess your existing software, systems or process at no charge. We quote the repairs and setup needed to work with what you already use. The subscription starts after we complete this work.
This is our target for small changes. Setup, complex work and waiting for your access or approval can take longer. If we need your input, the request stays active and the subscription continues.
You pay for hosting, domains and third-party services and licences. We quote brand and visual design separately. Our own AI costs are included.
You own the custom source code and documentation we create for you. We keep ownership of our development tools and reusable components.
You can cancel at any time. The subscription has no pause option. We finish the agreed active request and release it at no extra charge, even if the paid period has ended. If we are waiting for your approval or access after the paid period ends, we schedule the release when you respond and we have capacity.
We fix defects we cause at no extra charge. These fixes do not use your active request slot. After cancellation, this covers defects you report within 30 days of the final release. We offer no guaranteed response time.
In a 30-minute call, we discuss the software you need, the process you want to automate and where to start.