Automations
- Request and lead routing
- Data sync between tools
- Reports and alerts
- Document processing
- Support triage
Production-ready AI, agents and automation
I add AI features and agents to web applications and automate processes for startups and SMEs, fully remote. Each need is scoped first, because a plain automation sometimes does the job better than AI.
Jean-Baptiste Thery
Senior full-stack TypeScript & AI freelance developer
Agentic developer, within your rules
Available for fully remote freelance
- Starting point
- Scoping of one feature
- Pricing
- Fixed price, or day rate of €550 excl. VAT (indicative)
- Start date
- To be agreed
001
001·The service
Start from a need your users have
A need might be finding answers in long documents, routing incoming requests or sparing your team a repetitive task. I meet it with the simplest tool that works: a rule-based automation with n8n or custom code, an AI feature, or an agent acting within limits you set.
The model, the data it may use, the actions it may take and the interface then follow from your data constraints, since passages sent to an external AI service leave your infrastructure.
What the engagement covers
- A scoping note: the need, the data involved and the success criteria
- The feature or agent in your application, from the backend integration to the interface
- Automations with n8n or custom code where rules are enough, and AI only where they fall short
- Search with sources across your documents, or MCP connections to your tools, when the feature needs them
- An evaluation set of real examples with expected results, run before each release
- Limits on what the model may see, do and spend, with a confirmation before sensitive actions
- Documentation for your team
From a simple automation to an agent
AI features and agents
- Assistants and chat
- Search with sources
- Drafting from your data
- Agents with approvals
- MCP connections
002
002·How it works
From your first message to the handover
The same four steps on every engagement, so you know what happens next and what you receive.
You describe your project
Send a short brief on Malt or LinkedIn: your project, budget and target date. I confirm availability and the next step.
We agree the scope
Scope, schedule, pricing (fixed price or day rate), working hours and how we communicate, agreed in writing before work starts.
You follow the work
Code reviews, demos or written reports keep progress visible. Every change I deliver is reviewed and tested.
Handover and follow-up
Documentation, recommendations and, for development work, code whose rights the contract assigns to you. Maintenance by agreement.
003
003·Projects
AI products and features I have built
Interfaces for three AI startups, the AI coach of the product I co-founded and a search tool built for coding agents.
Adaptive AI coach of a workout generator
After every session, the coach adjusts each member's volume and rest from objective signals and suggests weights, with no generative model.
Cited search for coding agents
Coding agents such as Claude Code and Codex search your code and documents locally and get cited passages, through a CLI, a TypeScript API or an MCP server.
Dashboard for AI-generated questions and answers
A test-driven React dashboard to review and manage the questions and answers the AI produces, with PDF documents and several languages.
Frontend of an AI prospecting platform
Sales teams discover leads with AI, on the web and on mobile, through a React and Ionic frontend I built from scratch.
004
004·Recommendations
What clients and colleagues say
“I strongly recommend working with JB. He was instrumental in getting our frontend off the ground, handling everything with zeal, unwavering dedication, and unparalleled reliability.”
LinkedIn“Jean-Baptiste nous a aidé dans le développement frontend de notre outil Sweez. C'est un développeur talentueux et hautement compétent. Son expertise technique et sa capacité à résoudre les problèmes de manière créative font de lui un atout précieux pour toute équipe de développement. Il est fiable, consciencieux et démontre une grande attention aux détails dans son travail. Il communique bien et est force de proposition. La qualité de son code et sa capacité d'adaptation sont remarquables.”
Malt“Excellent niveau, s'acclimate vite avec son nouvel environnement de travail (compréhension de la stack, des outils de communication, de l'industrie), très autonome et même pro-actif sur la prise de décision. Excellente communication du travail en cours et restant. Se rend facilement disponible. Je recommande pour tout dev d'application web complexe et équipe haut niveau.”Malt005
005·About
AI features built like any other software
An AI feature is still software to design, test and maintain. I learnt that craft on projects such as Biodiv-Wind's real-time wildlife monitoring. That was before AI coding tools, and I bring the same discipline to AI work.
Through jCode, I work with public institutions, established companies and startups. My own products, WorkoutGen and Ragmir, taught me what follows a release.
- of web development
- 10 years
- missions completed on Malt (November 2026)
- 80+
- as a freelancer
- 7 years
- from 47 client ratings on Malt (November 2026)
- 5.0/5
Areas of expertise
006
006·Related services
Other ways to work with AI
Technical consultingGet advice on architecture and AI choicesRecoveryAudit an application built with AIAgent accessBuild an MCP server for your product
How I work with AI on client code is set out on the agentic development page.
007
007·FAQ
AI and automation for your business
Which AI features can you add?
Assistants and chat in your interface, search across your documents with sources, drafting from your own data, and agents that act through your API or MCP servers, with a confirmation before sensitive actions.
Is AI always the right answer?
No. Many processes need only a webhook, a scheduled job or an n8n workflow: cheaper, faster and easier to audit than a model. I suggest AI where rules fall short (unstructured text, classification, search), and an agent only when a task needs several steps and judgement.
Can you automate processes inside our company?
Yes. With your team, I map the process, its inputs and its exceptions, then automate it step by step with n8n or custom code, with logs and alerts that show what ran.
Will our data be sent to an AI provider?
Not necessarily. Your data constraints decide the model, the provider and what each request contains. The agentic development page sets out how I handle code and data.
How do you know the feature works?
We collect real examples with the results you expect and run them before each release. The results show where the feature helps, where it fails and whether a change made it better or worse.
Who should build it: a freelancer, an agency or a hire?
It depends on the scope and on who will own it: a freelancer for a defined feature in an existing product, an agency for several profiles in parallel, a hire for a core feature. The guide to this choice sets out the signs and what to check.
How much does it cost?
A clearly defined feature can be quoted at a fixed price after scoping; open-ended work is billed at an indicative €550 excluding VAT per day, as of October 2026. Model and hosting costs depend on usage, so scoping estimates them too.
008
008·Contact
Let's talk about your project.
Tell me what you want to build or improve, your budget and your target date.
- Fully remote
- English or French
- France · UK · US