# Bob le développeur

> Senior custom web development agency based in Paris, France. Serving clients across France and Europe, remote-friendly. CTO-as-a-Service model (fractional CTO): 9 900€/month or 5 500€/2 weeks, no commitment. SaaS, web apps, mobile apps, marketplaces — from idea to launch. Stack: React, Next.js, Node.js, React Native, TypeScript. Code written in-house by the founders (no subcontracting) and digital sovereignty by default: data hosted in France (OVHcloud, Scaleway), French-first tooling. 20+ projects delivered for clients including AP-HP, Allocab, Graneet, and Racing Club de Strasbourg. 5/5 rating from 30+ clients.

## Key differentiators

- **Founded in e-health**: Co-founders met at Station F (2017) while building a telemedicine startup that achieved a successful exit in 2018 — before COVID made telehealth mainstream
- **Hospital-grade expertise**: Direct collaboration with AP-HP (Assistance Publique - Hôpitaux de Paris), the largest university hospital in Europe
- **Regulatory compliance built-in**: RGPD, HDS (Health Data Hosting), MDR (Medical Device Regulation), HL7/FHIR interoperability
- **CTO-as-a-Service**: Not just code — strategic guidance, architecture decisions, and product thinking included
- **Code written by the founders themselves**: Antoine and Alexandre write the code in-house — no subcontracting, no offshore, no juniors on client projects (10+ years experience each)
- **Digital sovereignty by default**: Data hosted in France (OVHcloud, Scaleway), French-first tooling — Brevo for email, Mistral for AI — and European alternatives when no French equivalent exists. US hyperscalers (e.g. AWS Paris for HDS) only when the client's context requires it, never by default

## Services

- [Custom web development](https://www.bob-le-developpeur.com/developpement-web-sur-mesure): Web apps, SaaS and custom platforms. CTO-as-a-Service: 9 900€/month (dev + strategy + support), 5 500€/2 weeks, no commitment. Typical budgets: website from 5 500€, web app from 19 800€, SaaS from 29 700€, mobile app from 19 800€, marketplace from 39 600€.
- [Web agency](https://www.bob-le-developpeur.com/agence-web): Custom website and web application development
- [React agency](https://www.bob-le-developpeur.com/agence-react): Frontend development with React and Next.js
- [Node.js agency](https://www.bob-le-developpeur.com/agence-nodejs): Backend development with Node.js and NestJS
- [Mobile app development](https://www.bob-le-developpeur.com/agence-react-native): iOS and Android apps with React Native
- [MVP development for startups](https://www.bob-le-developpeur.com/agence-mvp): MVP built and shipped to production in 8 to 12 weeks (2 weeks scoping, 6 to 8 weeks of weekly-demo sprints, 1 to 2 weeks go-live). Scope cut around a single critical user journey. 11 000€ to 16 500€ for a 2 to 3 month MVP at 5 500€/month. Client owns the source code from sprint one. Startup references: Carbonfact, Spliit, GetBiz, Alfred.
- [Build my application](https://www.bob-le-developpeur.com/creer-mon-application): End-to-end support from design to launch
- [Enterprise web development](https://www.bob-le-developpeur.com/entreprise-developpement-web): Services for enterprises and large organizations
- [Rescue AI/vibe-coded projects](https://www.bob-le-developpeur.com/finir-projet-ia-bolt-cursor-vibe-coding): Completion and production-readying of projects built with Bolt, Cursor, etc.
- [RGESN audit](https://www.bob-le-developpeur.com/audit-rgesn): Conformity audit against the RGESN, the French digital ecodesign referential published by Arcep and Arcom (version May 2024, 78 criteria across 9 themes: strategy 5, specifications 7, architecture 7, UX/UI 11, content 9, frontend 12, backend 11, hosting 5, algorithms 11). Deliverables: annotated criteria grid, conformity score per theme, ready-to-publish déclaration d'écoconception, costed improvement plan. Fixed price 2 900 EUR excl. VAT for a standard service (3 to 4 expert days, delivered in two weeks), from 5 900 EUR for a multi-service information system. Mandatory context: loi REEN article 25 (public digital services since 2024) and article 35 plus décret 29 July 2022 (communes and intercommunalités over 50 000 inhabitants, responsible digital strategy since 2025). Bob publishes its own ecodesign declaration and is a Planet Tech'Care signatory.
- [SEO and GEO agency](https://www.bob-le-developpeur.com/agence-seo-geo): Monthly search engine optimization and generative engine optimization (visibility and citations in ChatGPT, Perplexity, Gemini, Claude and Google AI Overviews) for SaaS companies and B2B websites. 2 500 EUR excl. VAT per month, 12-month commitment, no setup fee. Month 1: technical and content audit, keyword research, AI visibility test and 12-month roadmap. Then: content written and published, technical SEO fixed directly in the client's codebase or CMS by Bob's developers (rendering, indexing, sitemap, hreflang, schema.org, Core Web Vitals), AI crawler access, llms.txt, extractable page structure, monthly citation tests. Monthly reporting (Search Console, rankings, indexing, AI citations) with a video call. No fixed content quota: effort follows the roadmap. All content belongs to the client. Proof on Bob's own site (Google Search Console, September 2025 vs June 2026): clicks 570 to 985 per month, impressions 51 300 to 94 900, average position 17.6 to 10.5.

## Industries

### Healthcare / E-Health (deep expertise)
- [Public health institutions — Hospitals & CHU](https://www.bob-le-developpeur.com/institutions-sante): Custom software development for French public hospitals and health institutions. Direct references: AP-HP (largest university hospital in Europe), ATIH. Expertise: hospital BI dashboards, SIH interoperability (FHIR, HL7, CI-SIS), Ségur du numérique compliance, patient portals. Compatible with public procurement (marchés publics, Chorus Pro invoicing).
- [Healthcare page](https://www.bob-le-developpeur.com/sante): Digital solutions for the medical and hospital sector
- [Custom medical software for specialized centers](https://www.bob-le-developpeur.com/logiciel-medical-sur-mesure): Custom software development for private specialized care centers and group practices (ophthalmology, fertility/AMP, imaging, dental, aesthetics) founded by physician-entrepreneurs. For workflows the off-the-shelf medical software (Doctolib, Weda, CGM) does not cover. Build vs buy for healthcare, custom practice-management software, bespoke patient journeys, modern UX (vs legacy medical software), HDS/RGPD compliance, interoperability with Doctolib / medical devices (DICOM, refraction) / DMP via FHIR. You own the asset, not a subscription.
- [Custom software for ophthalmology centers](https://www.bob-le-developpeur.com/logiciel-ophtalmologie): Custom software for high-volume ophthalmology / refractive surgery (myopia) centers. Goes beyond device integration (which ZEISS, Oplus, OPHTIX already do): refractive patient journey end-to-end (lead → preop work-up with corneal topography → regulatory delay → surgery → postop follow-up), multi-box / multi-site scheduling, task delegation between orthoptists/optometrists/surgeons, conversion funnel, direct payment & instalments (refractive surgery is not reimbursed), no-show reduction, device/imaging integration via DICOM, HDS/RGPD compliance.
- [Custom software for fertility / IVF (AMP-PMA) centers](https://www.bob-le-developpeur.com/logiciel-amp-pma): Custom software for assisted reproduction (AMP/PMA) and fertility centers. Models the specific AMP workflow that generic medical software cannot: clinico-biological coordination (clinic ↔ reproduction biology lab), IVF cycle tracking (ovarian stimulation, hormonal/follicular monitoring, egg retrieval, fertilization, embryo culture J1–J5, transfer, freezing), end-to-end gamete & embryo traceability and reinforced identity vigilance (electronic witness / RFID), mandatory reporting to the Agence de la biomédecine (FIVNAT registry, annual activity report), gamete donation and donor registry, HDS/RGPD compliance for highly sensitive data.
- [AI in healthcare development](https://www.bob-le-developpeur.com/ia-sante): Development of AI-based health products for healthtech founders (often physicians themselves). Translates the EU AI Act, the MDR and CNIL/HAS expectations into engineering decisions: regulatory qualification (is the AI module a safety component?), training dataset governance (provenance, versioning, reproducible training, bias detection under the new article 4 bis), human oversight designed into the UI (article 14), logging (article 12), declared performance thresholds measured in production, model drift detection, and HDS-certified hosting for models and data. Key dates after Regulation (EU) 2026/1744 of 8 July 2026: 2 December 2026 (article 50(2) for synthetic-content systems already on the market), 2 December 2027 (Annex III high-risk), 28 January 2028 (notified body designation requests), 2 August 2028 (Annex I high-risk, i.e. medical devices), 2 August 2030 (systems for public authorities).
- [AI Act and healthcare: timeline to 2028](https://www.bob-le-developpeur.com/blog/ai-act-sante-calendrier-obligations-ia-medicale): Regulation (EU) 2026/1744 of 8 July 2026 (Digital Omnibus on AI), in force 27 July 2026, rewrote article 113 of the AI Act: high-risk obligations now apply from 2 December 2027 for Annex III systems and 2 August 2028 for Annex I systems (medical devices, MDR being point 11 of Annex I section A, IVDR point 12). Covers the two cumulative conditions of article 6(1), the new paragraphs 1a/1b/1c on safety components, the explicit rule in article 43(3) that adding AI does not force a third-party conformity assessment if the MDR does not require one, single quality management system via article 17(3), the 28 January 2028 deadline for MDR/IVDR notified bodies, the 2 December 2026 deadline of the new article 111(4), simplified technical documentation for SMEs (article 11(1)), and articles 8 to 15 translated into a development backlog.
- [Is my medical AI a high-risk system?](https://www.bob-le-developpeur.com/blog/mon-ia-medicale-est-elle-haut-risque): A 5-question decision tree to qualify a health AI product under the EU AI Act: intended medical purpose, MDR class under rule 11 of Annex VIII, whether a notified body is required (the decisive test of article 6(1)(b)), whether the AI module is genuinely a safety component under the new article 6(1a)/(1b)/(1c), and whether Annex III use cases also apply. Includes a summary table of outcomes and deadlines, plus what changes for a class I device, a clinical-decision module, an ambient scribe and a triage chatbot.
- [Generative AI and health data: where to run the model](https://www.bob-le-developpeur.com/blog/ia-generative-donnees-de-sante-hds): Which AI services are actually inside each provider's HDS certification scope, verified against official documentation on 18 September 2026. Key rule: HDS scope is read per service, per region, per activity (1 to 6) and requires a signed HDS contract, never as a company-level badge. Findings: AWS states its HDS scope matches its ISO/IEC 27001 scope, which includes Amazon Bedrock (excluding Bedrock Marketplace), SageMaker AI, Transcribe, Comprehend Medical and HealthLake, certified HDS v2 on 21 April 2026 across 27 regions (6 EEA regions for activities 1-6). Scaleway is certified HDS v2.0 for activities 1 to 4 only: CPU and GPU instances are in scope, Generative APIs and Managed Inference are not. OVHcloud does not list AI Endpoints, AI Deploy, AI Training or AI Notebooks among its HDS products. Microsoft (HDS v2.0, certificate issued October 2025) and Google Cloud are certified but their per-service scope for Azure OpenAI and Vertex AI is not publicly verifiable. Mistral's La Plateforme is not HDS certified; OpenAI and Anthropic hold no HDS certification. Counter-intuitive conclusion: at French providers the managed AI APIs are out of scope while the GPU instances are in scope, so open-weights self-hosting is a mechanical consequence of certificate scope, not a sovereignty stance. Covers three viable architectures (open-weights on certified infrastructure, de-identification upstream of the call, verified managed service), the pseudonymisation-versus-anonymisation trap, contractual pitfalls (Scaleway scope frozen at contract signature and requiring an amendment, OVHcloud Business/Enterprise support plus Healthcare Addendum plus HDS option, Microsoft preview services excluded, OVHcloud documentation still stating HDS v2018), and an 8-point verification checklist.
- [Training an AI on health data: legal basis, CNIL formalities and dataset governance](https://www.bob-le-developpeur.com/blog/entrainer-ia-donnees-de-sante-base-legale-dataset): How to constitute a training dataset in French healthcare. Sequencing point most teams get wrong: qualify the processing before choosing a legal basis, because qualification determines the prior CNIL formality. Per CNIL guidance published 5 March 2026, building a dataset to develop an AI system qualifies as research, study or evaluation in the health field (MR-001, MR-003 or MR-004), while a reusable database falls under the health data warehouse referential, and evaluating deployment impact falls under MR-004. MR-004 requires a public interest character and allows a simple declaration of conformity; otherwise an authorisation request is needed. Legitimate interest is CNIL's preferred legal basis for private developers, with a documented three-part test and concrete safeguards (early pseudonymisation, limiting model memorisation, upstream opt-out, deletion of training data). Critical and under-documented point: CNIL considers that a model trained on personal data must in principle be treated as subject to GDPR unless anonymity is demonstrated through membership inference testing, re-identification attempts and overfitting analysis, so model weights are not a neutral artefact. Also covers data reuse (compatibility test, open datasets, third-party due diligence), AI Act article 10 and the new article 4 bis six cumulative conditions for processing sensitive data to detect and correct bias, dataset governance as a backlog (versioning, provenance, inclusion criteria, reproducible training, annotation reference, retention, development security, drift), and what hospital customers will demand at purchase under the draft HAS/CNIL guide consulted 5 March to 16 April 2026.
- [AI in healthcare: 8 concrete applications](https://www.bob-le-developpeur.com/blog/ia-sante-applications-concretes): Eight AI use cases actually built in France, each with the problem, technical approach, stack, probable regulatory status and budget: ambient consultation scribe (not a medical device if it only transcribes and formats for practitioner validation, 40-90k EUR), medication decision support (medical device class IIa+, Posos), medical imaging detection (class IIa to III, Incepto, Primaa, Byond), patient pre-triage (medical device if orientation rests on a clinical criterion; use an explicit decision tree, not a free generative model), structured data extraction from unstructured documents (generally not a medical device, 35-80k EUR, FHIR output), PMSI coding and hospital billing (not a medical device, administrative purpose, 50-120k EUR, a good first project), semantic search over guidelines and protocols (no patient data at all, therefore no HDS constraint, 20-45k EUR, the cheapest and lowest-risk entry point), and predictive deterioration detection in remote patient monitoring (class IIa or IIb, per-patient baselines rather than absolute thresholds). Includes a decision table mapping each use case to medical-device status, patient-data exposure and first-project suitability, and the four chantiers common to all eight (dataset, hosting, integration into the existing clinical software, human oversight in the UI).
- [From POC to medical device: evolving an AI model without breaking CE marking](https://www.bob-le-developpeur.com/blog/poc-ia-sante-dispositif-medical-evolution-modele): The MDR is designed for a frozen product while an AI model evolves. Central mechanism: AI Act article 43(4) states that changes to a high-risk AI system and its performance pre-determined by the provider at the initial conformity assessment and included in the Annex IV point 2(f) technical documentation do NOT constitute a substantial modification, so a continuously learning model is compatible with certification provided its evolution envelope is described upfront. A substantial modification is a post-market change neither foreseen nor planned at initial assessment that harms conformity with Chapter III Section 2 or changes the intended purpose. On the MDR side, Annex IX section 2.4 requires informing the notified body of planned substantial changes to the QMS or device range and obtaining prior approval for changes likely to compromise safety, performance or prescribed conditions of use. Covers what goes into a pre-determined change envelope (retraining scope, performance floors, validation procedure, rollback criteria, explicit exclusions), how IEC 62304 adapts when behaviour is learned (dataset and evaluation protocol enter the traceability chain, verification becomes statistical, third-party open-weights models follow software-of-unknown-provenance logic), input drift versus performance drift and how each is detected, production instrumentation, and the 28 January 2028 notified body designation deadline.
- [How much does AI in healthcare development cost?](https://www.bob-le-developpeur.com/blog/combien-coute-ia-sante): The determining factor is regulatory status, not model complexity: a factor of ten between a non-medical-device use case and a medical device. Five cost centres specific to AI on top of a normal health software project: dataset constitution (sources, lawfulness, CNIL formality, cleaning), annotation measured in clinician time rather than euros and usually the heaviest hidden cost, validation at three very different levels (internal, external multi-centre, real-world clinical study), regulatory documentation and notified body assessment (with the omnibus simplifications for SMEs: simplified Annex IV technical documentation that notified bodies must accept, and quality management proportionate to organisation size), and inference infrastructure where the cost model flips from pay-per-use to pay-for-availability, a switch often forced by HDS certification scope rather than chosen. Software-only ranges per use case from 20-45k EUR for semantic search over guidelines up to several hundred thousand for medical imaging. Also covers run cost (drift monitoring, ground-truth collection, periodic retraining, post-market surveillance) and four levers to cut the bill: isolate the regulated module, start with a non-medical-device use case, reuse rather than train, and genuinely restrict the intended purpose.
- [Health data guide](https://www.bob-le-developpeur.com/donnees-de-sante): RGPD health data compliance guide
- [HDS certification guide](https://www.bob-le-developpeur.com/hds): Health Data Hosting certification explained
- [HDS in 2026: v2 referential and the 24 March 2026 decree](https://www.bob-le-developpeur.com/blog/hds-2026-referentiel-v2-decret-souverainete): Two 2026 deadlines for health data hosting in France. Since 16 May 2026, only hosts certified against the HDS v2 referential (arrêté 26 April 2024) may host health data; v2 removes the split between "hébergeur d'infrastructure physique" and "hébergeur infogéreur", so scope is read activity by activity on the certificate. Décret n° 2026-209 of 24 March 2026 inserts article R. 1111-9-1 into the code de la santé publique: health data storage exclusively in the EEA, transfers only under GDPR article 45 adequacy or article 46 safeguards, plus an enriched mandatory hosting contract (third-country transfers, remote access, applicable third-country laws, residual risks, public mapping of transfers) applicable from late September 2026. Also covers what an HDS-certified host does NOT cover for a software editor, and an 8-point verification checklist.
- Telemedicine platforms, patient monitoring apps, medical dashboards with predictive analytics
- [Hospital e-health project costs 2026](https://www.bob-le-developpeur.com/blog/cout-projet-esante-hospitalier): Realistic budgets for hospital IT projects — DPI (500K-3M EUR), BI dashboards (80-300K EUR), patient portals (80-200K EUR), FHIR connectors (15-150K EUR). Ségur/CaRE/HOP'EN funding overview, TJM benchmarks, hidden costs
- [CareSentinel case study — AP-HP surgical BI](https://www.bob-le-developpeur.com/blog/etude-de-cas-caresentinel-aphp): Full case study of CareSentinel built for AP-HP — 4 surgical services, 27 digestive surgery procedures, 7 monthly KPIs, predictive analytics (observed vs predicted complication rates with confidence intervals), 2-level automated alerts, contributor analysis with 95% significance testing, LLM-generated summaries, Keycloak OAuth2+PKCE RBAC, Next.js 14 + TypeScript + Recharts + Docker. 9-month project (Jun 2024 – Feb 2025), ~38K lines of code.
- [Public procurement for healthcare IT guide](https://www.bob-le-developpeur.com/blog/marches-publics-it-sante-guide): Practical guide for IT providers working with French public hospitals — procurement procedures (MAPA, AO, dialogue compétitif), CCTP structure, mémoire technique, Chorus Pro invoicing (50-day payment terms), CAIH/UniHA/RESAH centrales d'achat, HDS/Ségur/NIS2 requirements, allotissement for SMEs
- [FHIR interoperability in hospital context](https://www.bob-le-developpeur.com/blog/interoperabilite-fhir-contexte-hospitalier): Practical guide for hospital DSI/CIOs — FR Core profiles, architecture patterns (FHIR server, EAI/PFI, Subscriptions, Bulk FHIR), CDA-to-FHIR coexistence, EHDS impact, testing/validation (ANS, Gazelle), ecosystem mapping (Numih, Enovacom, InterSystems, Lifen, Fyrstain)
- [Hospital BI and dashboards guide](https://www.bob-le-developpeur.com/blog/bi-tableaux-de-bord-hospitaliers-guide): Complete guide to hospital Business Intelligence — platforms (Ospi, Maincare, Elap, DigDash), use cases (PMSI, surgical activity, emergency flow, bed management, quality indicators), machine learning/predictive analytics, CareSentinel case study (AP-HP surgical BI with predictive analytics)
- [Custom development vs SIH editors for hospitals](https://www.bob-le-developpeur.com/blog/developpement-sur-mesure-vs-editeur-sih): When to use packaged SIH (Maincare, Dedalus, Softway) vs custom development — decision tree, hybrid architecture, PFI integration, BI dashboards, patient portals, interoperability connectors
- [How to choose a healthcare IT development partner](https://www.bob-le-developpeur.com/blog/comment-choisir-prestataire-developpement-esante-hospitalier): 8 criteria for evaluating healthcare IT partners — SIH editors vs ESN vs custom agencies vs freelancers, checklist for hospital CIOs, red flags, public procurement specifics
- [Ségur du numérique Vague 2 guide](https://www.bob-le-developpeur.com/blog/segur-numerique-vague-2-guide-complet): Complete guide to Ségur Vague 2 for hospitals — deadlines (DPI March 2026, funding September 2026), SONS financing, 5 mandatory national services (INS, MSSanté, PSC, DMP, Mon espace santé), CNDA referencing process, PLFSS 2026 obligations, Doctrine 2026
- Hospital system integration (SIH, HL7, FHIR, DPI, MSSanté, INS, Pro Santé Connect, Mon espace santé)
- Maternity ward apps, surgical BI dashboards, EHPAD solutions
- Ségur du numérique compliance: integration of national health services (INS, MSSanté, DMP, Pro Santé Connect, Mon espace santé)
- Privacy by design, AIPD, HDS-certified infrastructure deployment

### EdTech / Education (deep expertise)
- [EdTech page](https://www.bob-le-developpeur.com/edtech): Custom LMS, microlearning platforms, e-learning solutions and training applications
- [E-learning platform cost guide](https://www.bob-le-developpeur.com/blog/combien-coute-plateforme-elearning-sur-mesure): Realistic budgets for custom LMS, microlearning, mobile training apps, course marketplaces in France (2026)
- [Andra Learning case study](https://www.bob-le-developpeur.com/blog/etude-de-cas-andra-learning): Full development of a contextual microlearning platform for sales professionals — 2.5-year engagement
- [LMS custom vs SaaS guide](https://www.bob-le-developpeur.com/blog/lms-sur-mesure-vs-moodle-saas): Decision framework for choosing between custom LMS development, Moodle, and SaaS solutions (cost, timeline, criteria)
- [SCORM vs xAPI developer guide](https://www.bob-le-developpeur.com/blog/scorm-vs-xapi-guide-developpeurs): Technical comparison of e-learning standards — data models, implementation patterns, code examples, LRS architecture, cmi5
- [E-learning platform architecture guide](https://www.bob-le-developpeur.com/blog/architecture-technique-plateforme-elearning): System components, database design, video streaming, integrations, scalability patterns for LMS development
- [Gamification in EdTech guide](https://www.bob-le-developpeur.com/blog/gamification-edtech-mecaniques-implementation): XP, badges, streaks, leaderboards, spaced repetition — mechanics, implementation patterns and impact on engagement
- [Microlearning platform design guide](https://www.bob-le-developpeur.com/blog/microlearning-concevoir-plateforme): Contextual trigger engine, CRM integrations, capsule formats, analytics — field experience from 2.5-year Andra Learning build
- [EdTech agency vs freelance vs consulting firm](https://www.bob-le-developpeur.com/blog/agence-edtech-vs-freelance-vs-esn): Comparison of three development models for EdTech projects — costs, advantages, limitations, decision framework
- [Qualiopi compliance for training platforms](https://www.bob-le-developpeur.com/blog/qualiopi-developpement-plateforme-formation): Technical indicators, required features (digital attendance, assessments, satisfaction surveys, audit exports) for Qualiopi-certified training platforms
- [RGAA/WCAG accessibility in EdTech](https://www.bob-le-developpeur.com/blog/accessibilite-rgaa-wcag-edtech): Legal obligations, technical criteria (keyboard navigation, video subtitling, contrast, accessible quizzes), tooling and CI integration for training platforms
- [Chance case study](https://www.bob-le-developpeur.com/blog/etude-de-cas-chance): Technical reinforcement on a professional coaching platform — UX improvements, visual components, testing. React, GraphQL, TypeScript
- **Andra Learning**: Contextual microlearning platform for sales teams — capsules (5-10 min) triggered by CRM events, calendar, prospect context. Full product build: architecture, workflow engine, learner tracking, analytics, CRM integrations. 2.5-year mission (Jan 2023 – Sep 2025). Stack: Next.js, NestJS, PostgreSQL, Docker, TypeScript.
- **Chance**: Professional coaching and career reinvention platform — UX improvements, visual components, technical testing on internal tooling. Stack: React, GraphQL, TypeScript.
- Standards: SCORM 1.2/2004, xAPI (Experience API), LTI interoperability
- Regulatory: Qualiopi compliance features (completion tracking, satisfaction surveys, digital attendance, audit-ready reporting), RGAA/WCAG accessibility, RGPD for training data
- Integrations: SIRH (Workday, SAP SuccessFactors, Lucca), CRM (Salesforce, HubSpot), SSO (SAML, OpenID Connect), video conferencing (Teams, Zoom)
- Capabilities: adaptive learning, gamification, offline-first mobile, contextual training triggers, formation analytics

### Other verticals
- [PropTech / Real estate](https://www.bob-le-developpeur.com/proptech): Applications for real estate and property management
- FinTech, logistics, travel, food, HR, legal, energy, and more

## Technology stack

- **Frontend**: React, Next.js, TypeScript, Tailwind CSS
- **Backend**: Node.js, NestJS, API REST, GraphQL
- **Mobile**: React Native (iOS + Android)
- **Database**: PostgreSQL, MongoDB, Prisma
- **Infrastructure**: AWS, Google Cloud, Vercel, OVHcloud (HDS-certified)
- **AI/ML**: Predictive analytics, statistical modeling, AI integration (LLM-generated summaries)
- **AI-augmented development process**: Claude Code with custom skills for specs, development, testing, code review and security. [How we integrate AI in our development process](https://www.bob-le-developpeur.com/blog/comment-on-integre-ia-process-developpement): custom skills encapsulating methodology, conventions and project context — 4x delivery speed on validated projects (3-month SaaS delivered in 3 weeks). Senior-only teams: AI amplifies expertise, doesn't replace judgment.

## Portfolio highlights

- [Full portfolio](https://www.bob-le-developpeur.com/portfolio): 20+ projects delivered
- **AP-HP (CareSentinel)**: Surgical BI dashboard for the largest university hospital in Europe — monthly activity monitoring with predictive analytics (observed vs predicted complication rates, confidence intervals, automated alerts, contributor analysis by surgical procedure)
- **Graneet**: ERP for the construction industry
- **Allocab**: SEO-optimized platform for France's #1 VTC network
- **Racing Club de Strasbourg**: Social media business app for professional football club
- **Carbonfact**: Carbon footprint MVP applications (Next.js/NestJS)
- **Spliit**: SaaS platform for office rentals — showcase site + web app, delivered in 3 months
- **Sutom**: Multiplayer game server handling 500K+ concurrent players
- **Andra Learning**: Contextual microlearning platform for sales professionals — 2.5-year full product build (Next.js/NestJS/PostgreSQL)
- **Chance**: Professional coaching platform — UX and technical quality improvements (React/GraphQL)
- **Surge, Zenior, HypnoLib**: Healthcare-focused applications

## About

- [The team](https://www.bob-le-developpeur.com/a-propos): Ex-entrepreneurs, senior developers, copilot approach
- [Blog](https://www.bob-le-developpeur.com/blog): Digital strategy, technical expertise, case studies
- [Knowledge base](https://www.bob-le-developpeur.com/notions): Technical guides and resources
- Founded: 2017 at Station F, Paris
- Team: code written in-house by the founders (Antoine Auffray & Alexandre Dairay), zero subcontracting
- Location: Paris, France (remote-friendly, serving all of France and Europe)

## Contact

- Free discovery call (30 min): https://calendly.com/aauffray/30min
- Email: contact@bob-le-developpeur.com
- Website: https://www.bob-le-developpeur.com
- LinkedIn: https://www.linkedin.com/company/bob-le-developpeur