Daniel Neveux

Guiding a Product Owner from vibe coding to enterprise-grade delivery 𝗦𝗶𝘁𝘂𝗮𝘁𝗶𝗼𝗻 A Product Owner had just discovered AI-assisted development ("vibe coding"). He'd already grasped the key point — his technical teams could never match that pace — but ran into two walls: instability during development and security exposure. 𝗧𝗮𝘀𝗸 Preserve the velocity of AI-driven development while restoring the rigor and security his organization required, and give him something he could take to his managing director covering his entire application portfolio. 𝗔𝗰𝘁𝗶𝗼𝗻 Two fronts. On method, I walked him through applicable SDLC practices and coached him to get more out of Claude — segmenting the work, documenting it, versioning it. On diagnosis, I reverse-specified his codebase and produced a security report using Jared, my agentic pipeline: 467 reverse-engineered functional and business specifications and 997 code-to-spec links — the analytical backbone behind my Clever Spec offering. 𝗥𝗲𝘀𝘂𝗹𝘁 A professional security report (SOC 1 / SOC 2, OWASP), an exhaustive specification of the existing work, and a strong professionalization of what had started as a purely vibe-coded experience — delivered as a portfolio-wide plan ready to present to his DG.

https://www.instagram.com/ptidonjon/ 𝗦𝗶𝘁𝘂𝗮𝘁𝗶𝗼𝗻 A founder with a strong vision for an AI-powered editorial platform — personalized gamebooks for children. But no structured way to validate the market. No technical backbone to ship. 𝗧𝗮𝘀𝗸 Two missions: coach the founder through idea validation and product-market alignment, and build the technical platform behind what she validates. 𝗔𝗰𝘁𝗶𝗼𝗻 I coached the founder until she could run her own user interviews and pull real signal from them. One rule on the build: never write a line of code for something she couldn't sell first. We designed a minimal presale — enough product surface to sell the promise, without committing to the full generation engine or physical logistics. In parallel I built the technical platform: a multi-agent workflow that turns a short customer conversation into a personalized illustrated book. The architecture was shaped so the offer could pivot on the same backbone if the market signal pointed elsewhere. 𝗥𝗲𝘀𝘂𝗹𝘁 The founder runs the customer side independently. The minimum viable scope is shipped. The digital presale is live. A pivot scenario is already architected and ready. 𝗧𝗲𝗰𝗵𝗻𝗶𝗰𝗮𝗹 𝗱𝗲𝘁𝗮𝗶𝗹𝘀 Multi-agent workflow: Instagram chatbot → interview → brief → AI writers (OpenAI GPT) → AI illustrators (Replicate, Flux 2 + LoRA) → book assembly. NestJS micro-services (6+ feature modules). Event-driven + CQRS + SAM Pattern (own lib @warfog/ravioli, MobX). Payments: Stripe with webhook validation. Real-time tracking: Socket.io. Acquisition: Instagram Graph API. Auth: JWT, magic links, Passport.js.

𝗦𝗶𝘁𝘂𝗮𝘁𝗶𝗼𝗻 A construction marketplace (artisans / clients, BTP). Sales saturated by manual handling of incoming pricing requests. Tech was the bottleneck on every operational tweak. 𝗧𝗮𝘀𝗸 Automate the lead-to-CRM pipeline. Give Sales autonomy on the acquisition chain. 𝗔𝗰𝘁𝗶𝗼𝗻 v1: built an n8n workflow filtering inbound emails, classifying leads vs prospects, injecting into a custom CRM. Auto-reply B2B2C on pricing requests. Pipeline dashboard for Sales. v2: full rebuild powered by Jared. Acted as the operational bridge between Product, Tech, Sales, Ops on the whole acquisition chain. 𝗥𝗲𝘀𝘂𝗹𝘁 Sales decoupled from Tech on daily acquisition workflow. No more tech ticket for routine pipeline tweaks. 𝗧𝗲𝗰𝗵𝗻𝗶𝗰𝗮𝗹 𝗱𝗲𝘁𝗮𝗶𝗹𝘀 n8n workflows, custom CRM, B2B2C auto-reply automation. v2 rebuilt with Jared (Claude API, MCP).

𝗦𝗶𝘁𝘂𝗮𝘁𝗶𝗼𝗻 A VC-funded EdTech AI platform needed a new feature: school-specific content management with versioning and approval workflow. In-house estimate: 2 months for the existing team. 𝗧𝗮𝘀𝗸 Deliver the feature with Jared. Navigate the contractual constraints of a VC-funded scale-up. 𝗔𝗰𝘁𝗶𝗼𝗻 Negotiated the framework: IP clause (VC requirement), payment-before-delivery, scope trade-offs framed upfront (UI vs functional). Delivered the full functional scope: School Manager portal, multi-step Admin Review workflow (draft → pending → review → testing → published), Content versioning with side-by-side diff and rollback, Student view. Browser tests automated for epic validation. 𝗥𝗲𝘀𝘂𝗹𝘁 Delivered in 3 days vs 2 months estimated for the in-house team. x20. Client PO validation: "very broad scope coverage", "fully testable". 𝗧𝗲𝗰𝗵𝗻𝗶𝗰𝗮𝗹 𝗱𝗲𝘁𝗮𝗶𝗹𝘀 Prisma, NestJS, React, TypeScript, Firebase, Langfuse. Browser tests automated.

Technical Lead & Architect · L'Oréal face-tracking & photo-analysis SDK, white-label across 100+ brands. Focus: legacy modernization & AI-augmented delivery. 𝗣𝗵𝗼𝘁𝗼 𝗺𝗼𝗱𝘂𝗹𝗲: 𝗹𝗲𝗴𝗮𝗰𝘆 + 𝗔𝗜 𝗺𝗼𝗱𝗲𝗿𝗻𝗶𝘇𝗮𝘁𝗶𝗼𝗻 Situation: inherited photo module, untested, undocumented, bug-prone, coupled to a legacy service and moved to legacy. Action: reverse-engineered all specs; test coverage to 100% (unit, integration, E2E); synthetic-face scripts to regression-test the lighting algorithm; modernized the plugin with Jared, my agentic SDLC system. Result: higher stability, 100% coverage held, fewer support tickets. 𝗣𝗵𝗼𝘁𝗼-𝘃𝗶𝗮𝗯𝗶𝗹𝗶𝘁𝘆 𝗮𝗹𝗴𝗼𝗿𝗶𝘁𝗵𝗺 (𝗖 / 𝗢𝗽𝗲𝗻𝗖𝗩) Situation: low-level C/OpenCV algorithm (pixel measures, heuristics), kept identical client and server, with a buggy release. Action: aligned the Toronto teams on the reference version and fixed the algorithm. Result: higher photo reliability, shipped to production for millions of monthly users. 𝗔𝗴𝗲𝗻𝘁𝗶𝗰 𝗱𝗲𝗹𝗶𝘃𝗲𝗿𝘆 & 𝗹𝗲𝗴𝗮𝗰𝘆 𝗸𝗻𝗼𝘄𝗹𝗲𝗱𝗴𝗲 𝗿𝗲𝗰𝗼𝘃𝗲𝗿𝘆 (𝗝𝗮𝗿𝗲𝗱) Situation: a restructuring project was stalled after the chief engineers left, business logic left implicit. Action: built Jared, an agentic SDLC system, from my own needs; used it to re-engineer legacy knowledge into strict, exhaustive specs. Result: delivery gains of x5 to x10 in time and cost on my perimeter; project unblocked with a full spec set. 𝗗𝗲𝘀𝗶𝗴𝗻-𝘁𝗼-𝗰𝗼𝗱𝗲 (𝗙𝗶𝗴𝗺𝗮) Action: structured Figma with the design team and leveraged design-to-code through Jared. Result: screen integration under 5 min/screen at near-100% fidelity. 𝗔𝗱𝘃𝗶𝘀𝗼𝗿𝘆 SLA & quality governance: flagged the fragile points not to commit to. AI to top management: advised leadership on applied AI, from my Claude Code experience and SDLC acceleration on my perimeter. Stack: TypeScript, Node, React Native, Three.js, C/OpenCV, MediaPipe, Azure CI/CD, Claude/Claude Code, Jared.

Helped a blockchain startup to deliver a 3D web app to buy/sell parcels of their own metaverss project. The team was first in demand of feature implementation but when I audited the code it turned out that their was a strong technical debt due to strong coupling between the modules. If we have started with this code, I would not be able to deliver the features. Plus, there was a lack of project management skills which would lead to unfaisable project. I first helped the team to get a clear vision of their goals. I have implemented a Kaban board to cut down all the features and helped the team to decide. They showed abnegation and a great capacity for adaptation. Kudos to them. Main time, I have documented the way the code was working then did a first pass of refactoring. Once the code was cleaned and goals was clear, features were quickly implemented in all level (UX, shaders, etc.).

Helped a startup to release its first product in health, sport and IOT field. I was in charged of the React Native application. The team was reformed two times before. I first was contracted to implement React Native feature. It turned out there was a lack of project management methods, so I first implemented a cleaner flow under Jira for the client to get a better visibility of the developers achievements. One month later, I have been offered to lead the mobile team. My goal was to improve the developer experience to mount a scalable team while increasing the team velocity. As a result we finally managed to reach a rhythm of 2,5 releases per month. In a glance: - Led frontend architecture refactor to easily onboard junior developers. - Mounted and managed 4 members (international team). - Implemented workflow process for product team

Co-founded my first business at 23 yo. It was a visual communication agency specialized in emergent technologies. We worked for big accounts like : - Hermes - Suez - St Gobain With emergent technologies like : - Augmented reality (in 2010) - Virtual reality (in 2012) Won Awards : - Local innovation awards : CCI Morbihan - European awards : ESA Award