▄████▄
▄▄███████▄▄
▄█████▀▀▀▀▀▀█████▄
▄███▀ ▀███▄ ┌──────────────────────────────────────┐
████ FILIPPO ████ │ SOFTWARE ENGINEER │
▐████ ████▌ │ > [BACKEND] [GEN-AI] [MULTI-AGENT] │
████ VENTURINI ████ │ > [RESEARCHER] [OSS] │
▀███▄ ▄███▀ └──────────────────────────────────────┘
▀█████▄▄▄▄▄▄█████▀
▀▀███████▀▀
▀▀████▀▀
Software Engineer | AI Systems Architect | Open Source Contributor
Production backend systems · AI from research to production · Published researcher
class FilippoVenturini:
def __init__(self):
self.role = "Backend Engineer at Blue Reply"
self.impact = "Powering autonomous driving for 1M+ vehicles/day"
self.focus = ["LLM Agents", "Generative AI", "Multi-Agent RL"]
self.publications = ["GNN-based Multi-Agent RL paper"]
self.oss_contributions = ["CtxVault", "InvokeAI"]
def current_mission(self):
return {
"building": "Local RAG engines, multi-agent test pipelines, and DDPM from scratch",
"shipping": "Production backend systems for autonomous driving at 1M+ vehicles/day",
"researching": "Zero-shot scalability in multi-agent RL with GNNs"
}Multi-agent RL system coordinating drone swarms with zero-shot scalability via Graph Neural Networks.
PyTorch PyGeometric Gymnasium MARL · 10 → 29 agents · permutation-invariant policies
Offline-first semantic search engine for LLM pipelines. No cloud, no setup, agent-ready in seconds.
Python ChromaDB FastAPI · <100ms retrieval · published on PyPI
Full DDPM diffusion model rebuilt from scratch in PyTorch to understand how generative models actually work.
PyTorch DDPM UNet · custom noise schedule · custom variance optimization
3D asteroid impact simulator with real-time orbital dynamics. Evaluated by NASA mentors.
Node.js Three.js NASA NeoWs API · 60 FPS · 50+ NEO trajectories · <1% orbital error
[languages]
list = ["Python", "Java", "C#", "C++", "Scala", "JavaScript", "Bash"]
[ai_ml]
frameworks = ["PyTorch", "LangChain", "LangGraph"]
models = ["LLMs", "Diffusion Models (DDPM)", "MARL", "GNNs"]
rag = ["ChromaDB", "Embeddings", "Vector Search"]
[systems]
backends = ["FastAPI", "Spring Boot", "Node.js"]
databases = ["MongoDB", "MySQL"]
infra = ["AWS", "Kubernetes", "Docker"]
[frontend]
frameworks = ["React", "Vue.js", "Angular"]
[tooling]
list = ["Git", "GitLab CI/CD", "PyPI", "LaTeX", "Figma"]
[other]
list = ["Unity", "Arduino", "Embedded Systems"]| Degree | Institution | Grade | Years |
|---|---|---|---|
| BSc · Computer Science & Engineering | University of Bologna | 110/110 Honors | 2019–2022 |
| MSc · Computer Science & Engineering | University of Bologna | 110/110 Honors | 2022–2024 |
| MSc · Artificial Intelligence | Polytechnic Univ. of Catalonia | Erasmus | 2023–2024 |