LegalTech / GovTech
01
BESCHEIDKLAR
AI-assisted LegalTech SaaS platform
Own product04/2025 – today · Founder & CTO
Role
Founder & CTO
Team
Solo · own product (side project)
Location
Lower Saxony, DE · remote
Language
German
Website
Methodology
Direct product ownership
Responsibilities
- Built a modular, service-oriented architecture for the claimant-facing frontend, the law firm area and the regional licensing model (Stripe).
- AI pre-assessment pipeline for official notices and terminations (in progress): extraction → deadline check → plausibility and prospects assessment, with the case prepared in structured form.
- Automated, GDPR-compliant handover of qualified leads including uploaded documents to matching specialists (lawyers, law clinics).
- RAG-based guidance section with articles and recommended next steps for those affected.
- Coverage of several legal areas (civil, social and employment law) through an extensible, multilingual and low-barrier web interface.
- End-to-end responsibility for the tech stack, hosting on EU infrastructure and a secure-by-design compliance concept (EU AI Act, GDPR, German Legal Services Act).
- Type-safe end-to-end data flows (TypeScript, DTOs, Zod) with automated CI/CD delivery (GitHub Actions, Docker), monitoring and error tracking (Datadog) and test coverage (Vitest, Jest, Playwright).
- Close collaboration in code reviews and use of AI-assisted tools (Claude Code, Cursor AI, Code Rabbit) to safeguard quality and development pace.
Outcome
- MVP delivered single-handedly — architecture, platform and operation on EU infrastructure from one source. The remaining scope is in progress.
- Compliance built into the product design rather than retrofitted: the EU AI Act, the GDPR and the limits of the German Legal Services Act were part of the plan from the start.
- Delivery fully automated: tests, build and container deployment run through GitHub Actions, and production errors surface through Datadog — releases with no manual step, even as a one-person team.
Targets
- Up to 10,000 automated pre-assessments per month and up to 80 % less manual case intake.
- A predictable inflow of pre-qualified cases for partner law firms instead of costly manual initial assessment.
Technologies
TypeScriptJavaScriptPythonSQLBashHTMLCSSJSONYAMLUMLNext.jsReactServer-Side Rendering (SSR)FastAPIBackend developmentFrontend developmentWeb developmentArtificial intelligenceGenAIAI EngineeringLLM integrationOpenAI APIAzure OpenAIRAG (retrieval augmented generation)Prompt EngineeringEmbeddingsVector databasespgvectorSemantic searchOCRDocument analysisLangChainAI agentsMulti-agent workflowsAI-assisted automationAI-assisted test generationGDPR-compliant AI architectureClaude CodeClaude Code HooksSlash CommandsAgent SkillsSubagentsChatGPTv0Agentic CodingAgentic Software EngineeringContext EngineeringSpec-Driven DevelopmentPostgreSQLRedisData architectureTailwind CSSshadcn/uiComponent librariesStorybookDesign TokensWireframingInternationalisation (i18n)Accessibility (WCAG)German Accessibility Reinforcement Act (BFSG)European Accessibility Act (EAA)StripePayment processingSEPAJWTOAuth 2.0Secure by DesignIT securityEnd-to-end encryptionDockerGitHub ActionsCI/CDContinuous IntegrationDevOpsHetznerIONOSEU hostingLinuxUbuntuDatadogMonitoringObservabilityVitestJestPlaywrightUnit TestingIntegration testsEnd-to-end testsAPI testingPostmanTest automationZodDTOsSlackHuskypnpmCLI (Command Line Interface)Software architectureAPI-FirstMiddlewareDesign PatternsDomain-Driven DesignMicroservicesScalable architectureClean CodeSOLIDClean ArchitectureMVP developmentSaaSLaw firm softwareProcess digitalisationEU AI ActAI GovernanceGDPRPrivacy by DesignData minimisationData processing agreement (DPA)Digital sovereigntyKanbanProject managementIT strategyDigital strategyDigital transformationSoftware developmentProduct ownershipEntrepreneurshipInterface developmentAPI integrationLegalTechGovTech