in produzione 24/7in production 24/7
~/systems/ai-gateway
Gateway LLM multi-providerMulti-provider LLM gateway
Gateway self-hosted OpenAI-compatible: routing adattivo con Thompson sampling su 13 provider e 163 deployment di modelli, hedging della tail-latency, rate limiting GCRA, circuit breaker a 3 stati con auto-heal, key-pool con rotazione cooldown-aware. Osservabilità completa: 15+ metriche Prometheus, OpenTelemetry, SLO burn-rate multi-window, alerting deduplicato.
Self-hosted OpenAI-compatible gateway: adaptive routing via Thompson sampling across 13 providers and 163 model deployments, tail-latency hedging, GCRA rate limiting, 3-state circuit breaker with auto-heal, cooldown-aware key-pool rotation. Full observability: 15+ Prometheus metrics, OpenTelemetry, multi-window SLO burn-rate, deduplicated alerting.
235
file di testtest files
Personal AI Knowledge SystemPersonal AI Knowledge System
R&D
"Secondo cervello": knowledge graph temporale + retrieval ibrido a 3 lane (denso + BM25 + graph walk, fusione RRF) con variante HippoRAG-2 su Personalized PageRank. Agente proattivo con confidence-routing e critic anti-contraddizione."Second brain": temporal knowledge graph + 3-lane hybrid retrieval (dense + BM25 + graph walk, RRF fusion) with a HippoRAG-2 variant on Personalized PageRank. Proactive agent with confidence-routing and an anti-contradiction critic.
1.394 test5 output channelRAG · KG
Lead-Gen B2B per startupB2B Lead-Gen for a startup
startup TorinoTurin startup
Engagement commerciale (startup torinese di installazioni immersive): scraping multi-fonte, enrichment contatti, scoring deterministico a 100 punti, CRM self-hosted con pipeline Kanban, outreach + go-to-market a 90 giorni.Commercial engagement (Turin immersive-installation startup): multi-source scraping, contact enrichment, deterministic 100-point scoring, self-hosted CRM with Kanban pipeline, outreach + 90-day go-to-market.
scoring 100ptCRM 6 entitàCRM 6 entitiesB2B sales
Quantitative Finance PlatformQuantitative Finance Platform
produzioneproduction
Piattaforma full-stack di finanza personale: 9 engine analitici schedulati (forecast, anomaly, Monte Carlo sulla probabilità di rovina), 4 feature LLM con circuit breaker (chat text-to-SQL sandboxed), security rewrite documentata (WebAuthn, CSRF, CSP).Full-stack personal-finance platform: 9 scheduled analytics engines (forecast, anomaly, Monte Carlo ruin probability), 4 LLM features behind a circuit breaker (sandboxed text-to-SQL chat), documented security rewrite (WebAuthn, CSRF, CSP).
1.160+ test183 endpointReact 19 · FastAPI
Content Intelligence PipelineContent Intelligence Pipeline
produzioneproduction
Pipeline distribuita su 2 nodi: 20 collector multi-fonte con classificazione LLM end-to-end, delivery crash-safe via WAL file-based, dedup multi-layer (URL, titolo cross-source, fingerprint), hand-off strutturato verso il knowledge graph.2-node distributed pipeline: 20 multi-source collectors with end-to-end LLM classification, crash-safe WAL-based delivery, multi-layer dedup (URL, cross-source title, fingerprint), structured hand-off into the knowledge graph.
894 test12.500+ item2 nodi ARM64
Resale-Intelligence PlatformResale-Intelligence Platform
produzioneproduction
Monitoraggio e qualificazione automatica di offerte second-hand: coda LLM durevole zero-loss, qualificazione vision multi-provider con failover, parser multilingua (5.800+ nomi), scorer a convergenza di 6 segnali con verdetto ternario.Automated monitoring and qualification of second-hand deals: durable zero-loss LLM queue, multi-provider vision qualification with failover, multilingual parser (5,800+ names), 6-signal convergence scorer with ternary verdict.
2.255 test~5.000 verdetti/g~5,000 verdicts/dayVLM · failover
Data-Extraction InfrastructureData-Extraction Infrastructure
produzioneproduction
Servizio condiviso di browser-automation engine-pluggable: 6 backend di rendering dietro un'unica API con fallback a caldo, motore anti-bot YAML-driven (8 segnali, 31 predicati), sessioni "human-shaped" — consumato da 3+ progetti.Shared engine-pluggable browser-automation service: 6 rendering backends behind one API with hot fallback, YAML-driven anti-bot engine (8 signals, 31 predicates), human-shaped sessions — consumed by 3+ projects.
313 test6 engineinfra condivisa