Building the future of AI-driven systems | 6+ years of production ML experience | $1.5M+ in proven cost savings

I'm a visionary AI/ML Leader and Solution Architect with expertise in Large Language Models, Multi-Agent Systems, and MLOps at enterprise scale. From revolutionizing aerospace with AI optimizations to architecting autonomous agents for procurement, I've led cross-functional teams to deliver breakthrough innovations that drive multi-million-dollar efficiencies and reshape industries. Passionate about building the future of AI, I thrive in high-growth startups and big tech environments, turning ambitious ideas into scalable realities.

Current Focus: AI architecture at askLio (YC23), engineering the world's first multi-agent system for procurement to automate complex workflows and unlock unprecedented efficiency.

• Pioneered $1.5M cost reduction in rocket manufacturing via AI-driven optimizations, setting new industry benchmarks
• Boosted material science predictions by 9.2% with cutting-edge GNNs, accelerating R&D cycles
• Reduced experimental efforts by 15% through innovative ML architectures, published at NeurIPS
• Garnered 200+ GitHub stars for open-source tools adopted by enterprises worldwide
• Led 20% efficiency gains in multi-million euro ad budgets using RL + LLM hybrids

TypeScript architecture testing framework inspired by Java ArchUnit

• 200+ stars | Enforces architectural rules and dependency constraints
• Impact: Adopted by enterprises for maintaining clean code architecture

Domain-specific BERT for engineering literature

• Tech: PyTorch, Transformers, Custom Tokenization
• Impact: Specialized NLP model for technical documentation

AI-driven rocket design optimization

• Impact: $1.5M cost reduction, 15% improved payload/cost ratio
• Tech: GNNs, Neo4j, trajectory optimization algorithms

• Architecting the world's first multi-agent system for procurement, automating end-to-end workflows for global enterprises
• Leading AI strategy to integrate LLMs and RL for intelligent decision-making, targeting 50%+ efficiency gains in supply chain operations
• Collaborating with YC-backed teams to scale innovative AI solutions from prototype to production, driving startup growth

• Built AI-powered support solutions improving voice chatbot latency by 30% and onboarding time by 70%
• Led scalable AI infrastructure serving thousands of interactions for enterprise clients
• Developed autonomous prompt agent boosting satisfaction by 34% and reducing escalations by 45%

• Architecting RL models with LLM-driven analytics for multi-million euro ad budgets
• Engineered autonomous AI agents for real-time data synthesis and campaign optimization
• Delivered 20% efficiency improvement in ad spend with 15% ROI increase

• Advanced GNN development achieving 9.2% accuracy improvement in material predictions
• Multi-task transformer architectures with 4.7% representation accuracy gains
• Published at NeurIPS 2024 - cutting experimental effort by 15%

• Fine-tuned 70B LLM (DeepSeek R1) with LoRA achieving +4% BLEU and +6% ROUGE-L on benchmarks
• Implemented production RAG system with 92% retrieval accuracy on 100K+ documents
• Optimized inference strategies for 3x throughput and comprehensive evaluation frameworks

• Led team of 3 reporting directly to COO on AI-driven rocket optimization
• Architected Neo4j graph database for GNN-based design generation (5000+ configurations)
• Delivered $1.5M cost reduction through AI-optimized materials and structures

• Automated FEM analysis pipeline using Gaussian Processes, reducing cycle time by 25%
• Developed LSTM anomaly detection for satellite data with interpretable attention mechanisms

• Led mobile banking app development serving 20K+ users, presented at "Die Höhle der Löwen"
• Implemented ML models for transaction categorization and fraud detection on 1M+ records
• Managed 10-developer team establishing real-time inference pipelines

• Built job recommendation system facilitating 100+ internal transitions via collaborative filtering
• Implemented document classification achieving 89% F1-score on 50K documents

# LLM & GenAI
- Frameworks: LangChain, Hugging Face (Transformers, PEFT), OpenAI API
- Fine-tuning: QLoRA, Distillation, RAG Systems
- Vector DBs: ChromaDB, Qdrant, Neo4j
- Agentic Systems: Multi-agent frameworks, autonomous decision-making

# Core ML
- Deep Learning: PyTorch, TensorFlow, Advanced architectures
- Specialized: GNNs, Reinforcement Learning, Multi-task learning
- Optimization: Model compression, quantization, ONNX, TensorRT

# MLOps Pipeline
- Orchestration: ZenML, Airflow, Kubeflow
- Tracking: MLflow, W&B, CometML
- Cloud: AWS SageMaker, Azure ML, GCP Vertex AI

# DevOps
- Containers: Docker, Kubernetes
- Cloud: AWS, Azure, GCP (multi-cloud expertise)
- Monitoring: Prometheus, advanced observability

"Reaction Graph Networks for Inorganic Synthesis Condition Prediction"

• NeurIPS 2024 | 15% reduction in experimental effort
• Novel GNN architecture for materials science applications

"Aerospace-Specific LLM for Enhanced Research Summarization"

• 60% computational cost reduction
• Domain-specialized language models for technical literature

• World-Leading Multi-Agent Systems: Revolutionizing procurement and enterprise automation with scalable, intelligent agents
• Distributed ML Platforms: High-performance MLOps for real-time, global-scale deployments
• Domain-Specific AI Innovations: Custom LLMs and agents for high-stakes industries like aerospace and finance

• Architecture Testing Tools: Expanding ArchUnit-TS ecosystem
• MLOps Utilities: Production-ready tools for ML lifecycle management
• Educational Content: Tutorials and workshops on advanced ML techniques

• Scaling AI at innovative startups like Cursor or consulting firms
• Leading AI initiatives in big tech to drive global impact

Ready to collaborate on game-changing AI? Let's discuss:

• AI/ML Leadership for production systems and startups
• Research & Innovation Partnerships
• Opportunities in High-Growth Tech (startups, consulting, big tech)
• Open Source Contributions to ML tooling

📧 Reach out: jan@heimann.ai | 🌐 Portfolio: heimann.ai

"Engineering AI that transforms industries and creates lasting value."

⭐ If you find my work valuable, consider starring my repositories!