AI Systems
AI consulting, agent architecture and ML implementation for measurable workflow impact.
Freelancer: Build + Creator Package
From technical architecture to digital media execution: AI workflows, scalable product builds, video, sound and creator-led storytelling.
Voicfy Platform and Real Voice TTS
Agentic operations systems with audit trails
Film score and campaign sound delivery
Art direction and digital media production
AI consulting, agent architecture and ML implementation for measurable workflow impact.
UX, product architecture and implementation from MVP to scalable platform features.
Voice production, film scoring and sonic identity for campaigns and digital products.
Creative direction across digital media, motion, video and audio production pipelines.
Real Delivery
Voicfy
AI Voice SaaS Platform
Built a full-stack voice platform from product concept to production deployment, including billing, user flows and scalable architecture.
Created a reusable platform foundation for rapid feature rollout and monetization.
Voicfy
Voice Model Engineering
Developed a diffusion-based real voice TTS system with training and inference pipelines for production usage.
Improved output quality and enabled reliable high-volume voice generation workflows.
Treppe Aufwaerts (Feature Film)
Feature Film Score and Sound Direction
Composed and produced the original score for Treppe Aufwaerts, a German feature drama (2015) by Mia Maariel Meyer. The music direction was built around scene tension, family conflict and editorial rhythm, from thematic sketching to final mix delivery.
Established a coherent sonic identity across key character arcs and supported emotional continuity between dialogue-driven and high-pressure sequences. The score approach helped connect narrative pacing from festival presentation to theatrical release in Germany.
KaDeWe Group GmbH
Retail Digital Experience Design
Designed digital experience assets and visual systems for KaDeWe Group brand and campaign touchpoints.
Improved visual consistency across premium retail communication formats.
UX systems, interface direction and conversion-focused product experiences.
Proof: Product and brand systems for SaaS teams
Motion assets for product launches, campaigns and social formats.
Proof: Campaign motion packages with faster delivery cycles
Voice and sound identity that strengthens storytelling and brand recall.
Proof: Film scoring and commercial voice delivery
From AI strategy to agent workflows and production ML implementation.
Proof: Agentic Operations Copilot
Selected work across engineering, AI systems and creator-focused art direction.
Built a full-stack AI voice platform from product concept to production deployment.
Implemented agentic triage, routing and execution with human-in-the-loop controls.
Delivered original film score and sound design aligned with editorial pacing.
Directed digital visuals and campaign design assets for KaDeWe Group premium retail communication.
From strategy to production delivery: AI roadmaps, agent systems and ML models with measurable business impact.
Use-case prioritization, data and tooling audit, risk analysis, clear pilot scope.
Fast implementation with real data, evaluation, guardrails and team feedback.
Production integration, monitoring, governance and enablement for internal teams.
Problem: Service teams handled triage manually across email, Notion and CRM.
System: Agentic routing with human-in-the-loop for prioritization and escalation.
Stack: OpenAI · n8n · HubSpot API · Notion API
Result: Stable workflow for intake, classification and follow-up.
KPI: -38% manual handling time
Problem: Unqualified leads consumed sales capacity in first-contact stages.
System: Lead scoring, response drafting and scheduling logic with guardrails.
Stack: OpenAI · Pipedrive API · Slack · Webhook Orchestration
Result: Higher qualification quality before sales calls and cleaner handoff.
KPI: +24% SQL rate
Problem: Creators and teams needed more natural and controllable AI voices for production use.
System: Built an AI Real Voice TTS based on a diffusion model, including training and inference pipelines.
Stack: Python · PyTorch · Diffusion Model · FastAPI · Docker
Result: Production TTS system for Voicfy with improved voice quality and robust deployment.
KPI: Lower latency and significantly more natural voice output
Problem: Manual news evaluation was too slow and inconsistent for timely stock decisions.
System: Platform that ingests current news, normalizes signals and generates per-stock fundamental analysis.
Stack: Python · NLP Pipelines · LLM Analysis Layer · Market Data APIs · Dashboard UI
Result: Faster research workflow with structured fundamental assessment per stock.
KPI: Significantly reduced analysis time per stock
LLM orchestration | Vector DB / retrieval | Tool/API integrations | Evals and quality gates | Observability and tracing | Model serving and monitoring
Input -> Router Agent -> Tool Agents -> Human Approval -> Action + Log
Data -> Features -> Train -> Evaluate -> Serve API -> Monitor -> Retrain
Recovered media and legacy references that document long-term delivery across design, motion and sound.
Product interfaces, branding systems and practical UX implementation.




Historical case assets that support the current freelancer portfolio.




Campaign videos and digital motion pieces archived from the old site.
Artutopia Ad
Facebook Title Video
App Motion Cut
CW Motion Piece
Xinemate Teaser
Thex Motion
Film-score and voice branding references with visual covers.

Original score work for independent film projects and emotional narrative scenes.

German voice over and spoken-word branding work with commercial tone and character.
Founder context and external profiles for additional due diligence.