About
Hi, I’m Gian Paolo Santopaolo—welcome to my blog!
I’m a Senior Generative AI Engineer focused on building production-grade AI systems—from custom neural architectures (CNNs, DNNs, Transformers) and LLM/diffusion model fine-tuning to RAG platforms and agentic pipelines. I work daily with Python and PyTorch, delivering AI solutions from research prototypes to scalable, multi-GPU deployments.
I developed CogniX, an enterprise RAG platform and multimodal AI stack, and integrated generative AI (LLMs, diffusion, vision, speech, OCR) into Collaboard, a real-time collaboration product used worldwide.
I’ve designed and deployed into production multiple agentic and generative AI systems, including ReforgeAI (legacy code modernization), Sentinel-AI (event monitoring), CreativeCampaign-Agent (marketing automation), and DeltaE (color correction for fashion AI).
Future Tech 2019 - Amsterdam
Professional Highlights
- 24+ years in software engineering and architecture
- Deep focus on generative AI since 2017—before it was mainstream
- Microsoft MVP (2012–2022) and Microsoft Regional Director (2018–2020)
- Built AI platforms serving 100+ million users
- Led teams from monolithic to microservices migration at scale
- Inventor of real-time whiteboard technology that influenced Miro and Mural
Core Technical Expertise
Deep Learning & Generative AI
- PyTorch expert with daily hands-on coding in production environments
- LLM fine-tuning (Pixtral multimodal text+vision, GPT models, embedding models)
- Diffusion models (Flux, Stable Diffusion) including fine-tuning and style transfer
- RAG systems, knowledge graphs, and semantic search at scale
- Distributed inference on multi-GPU/multi-node clusters (vLLM, Hugging Face TGI)
- Multimodal AI: DeepSeek-OCR, SmolDocling, Tesseract, YOLO (custom-trained), Whisper
Agentic AI & Orchestration
- Production experience with CrewAI, Microsoft AutoGen, Semantic Kernel, Langflow, Google Antigravity
- LangChain & LangGraph for RAG and multi-step workflows
- Agent evaluation with Inspect AI and LangTrace
Model Serving & Evaluation
- FastAPI for model-serving APIs, Streamlit/Gradio for interactive UIs
- Experiment tracking with Weights & Biases, TensorBoard
- Observability with Grafana/Prometheus/Loki, Azure App Insights, CloudWatch, Google Cloud Monitoring
Cloud & GPU Infrastructure
- Bare-metal Kubernetes with NVIDIA Container Toolkit for GPU workloads
- Azure (strong): AKS, App Service, Functions, Cosmos DB, Azure OpenAI, AI Studio, Azure ML, Cognitive Services
- AWS: EC2, EKS, Lambda, S3, SageMaker, Bedrock, Comprehend, Rekognition
- GCP: GKE, Cloud Run, Vertex AI, Vision AI, Natural Language AI
- Lambda Labs/Cloud, RunPod for GPU compute
Software Engineering
- Python (expert), .NET/C# (expert), Go
- Modular design, API boundaries, testing, code review discipline
- Hydra-based configs for reproducibility and environment scoping
- Modern Python tooling: uv, just, pydantic
What I Build
I specialize in taking AI from research prototypes to production-grade systems:
✅ Event-driven microservices for AI workloads
✅ Multi-GPU distributed inference at scale
✅ RAG platforms with vector search (Qdrant, Milvus)
✅ Agentic systems using CrewAI, LangChain, LangGraph
✅ Model fine-tuning pipelines with proper evaluation
✅ End-to-end observability for AI services
I’ve shipped AI features to millions of users and built open-source projects that demonstrate production engineering patterns.
Awards & Recognition
- Microsoft Regional Director (2018–2020) – Selected among ~140 experts worldwide for technical excellence, business acumen, and community leadership
- Microsoft Most Valuable Professional (MVP) (2012–2022) – Recognized for deep technical expertise and community contributions
- Innovation Recognition – Inventor of real-time whiteboard technology that influenced platforms like Miro and Mural, today commercialized as Collaboard
Selected Speaking Engagements
- Stanford AI Professional Certificate Program – Show and Tell Sessions (October 2025: CV-Pilot and Fine-Tuning; September 2025: Reforge AI and Sentinel AI – 200+ attendees each)
- API World, San Jose (2023) – “Our Journey from Monolithic to Microservice with Kubernetes” (300+ developers)
- Future Tech, Amsterdam (2019) – Multiple AI/ML sessions (500+ attendees)
- .NET Conference (2018) – “AI for Every Developer” (3,000+ online)
- DevSum, Stockholm (2018) – “Machine Learning for Developers”
- ESPC, Copenhagen (2018) – “Machine Learning for Developers”
- Insider Dev Tour, Zurich & Milan (2018) – Keynote speaker
Education & Certifications
Stanford Artificial Intelligence Professional Program
- CS224N: Natural Language Processing with Deep Learning
- CS234: Reinforcement Learning
- CS236: Deep Generative Models
Machine Learning Specialization (DeepLearning.AI & Stanford University)
- Supervised Machine Learning: Regression and Classification
- Advanced Learning Algorithms
- Unsupervised Learning, Recommenders, Reinforcement Learning
Mathematics for Machine Learning Specialization (Imperial College London)
- Linear Algebra
- Multivariate Calculus
- Principal Component Analysis (PCA)
B.Sc. Applied Artificial Intelligence - IU International University of Applied Sciences
- 180 ECTS credits
- Currently enrolling
Continuous professional development through hands-on projects, research, and industry engagement.
Connect With Me
📧 Email: gianpaolo.santopaolo {at} gmail.com
📱 Phone: +41 78 340 32 28 | WhatsApp | Signal
💻 GitHub: github.com/gsantopaolo
🔗 CogniX: github.com/gen-mind/cognix
Need Help with Your AI Project?
Whether you’re building a new AI solution, scaling an existing system, or training your team, I can help. Book a free consultation to discuss your needs.
I look forward to connecting with you!