Giovanni Beijl

Cortex, a standalone AI infrastructure product · The Canopy Corporation

• Designed and built Cortex as a standalone AI infrastructure platform under The Canopy Corporation
• Built multi-tenant AI platform architecture for knowledge bases, assistants, agents, conversations, documents, scoped API keys, service accounts, provider configuration, usage analytics, and audit logs
• Designed platform scoping around organizations, workspaces, projects, environments, customer-owned resources, and platform-owned resources
• Built RAG workflows including knowledge-base retrieval, one-shot Q&A, persistent conversations, citations, feedback capture, retrieval scoring, and streaming Server-Sent Events responses
• Developed ingestion pipelines for JSONL, PDFs, codebases, documentation libraries, sync sources, embedding backfills, reindexing, cleanup jobs, and asynchronous ingestion job tracking
• Built assistant and agent configuration workflows, OpenAI-compatible endpoints, model/provider selection, system prompts, retrieval settings, tool configuration, and scoped API access
• Built model training workflows for QLoRA, LoRA, and full fine-tuning, including dataset export/formatting, tokenizer-aware truncation, adapter-aware inference, lightweight evaluation metrics, training-run tracking, diagnostics, and hardware snapshots

Built solo · canopy.pm

• Designed and built an EU-first developer cloud platform for deploying traditional applications and AI-oriented workloads
• Built repo-to-production deployment workflows where applications deploy directly through repository updates
• Built automatic project configuration, custom deployment pipelines, infrastructure orchestration, database provisioning, SSL/domain handling, and developer workflows
• Created one-click database provisioning inside project networks, configurable Nginx handling, and custom deployment configuration parsing through config.canopy
• Built a CLI tool for managing projects and deployment workflows, with support for Dockerfile and Docker Compose based deployment flows
• Owned backend, infrastructure, system design, deployment automation, operational reliability, and product trade-offs end-to-end

• Built AI-powered financial analytics functionality for an enterprise client, enabling PowerBI-style revenue, cost, profitability, quarterly comparison, and location-based insights through natural-language queries and automated chart generation
• Helped deliver deal-critical analytics and AI functionality that contributed to a multi-million-euro commercial opportunity
• Designed and built production-grade queueing infrastructure for heavy and delayed background jobs, including priorities, deduplication, retries, delayed scheduling, failure logging, CLI monitoring, and correlation-based batch progress
• Implemented analytics infrastructure with tenant-aware CLI tooling for relational-database-to-OLAP synchronization, health checks, schema drift repair, validation, and controlled cron scheduling
• Built agent-oriented tooling for autonomous knowledge management, including correction classification, canonical rule consolidation, audit logs, route synchronization, validation/linting, memory maintenance, and autonomous commit protocols
• Built public-sector identity integrations combining OIDC, SOAP-based government services, mutual TLS certificates, RSA-SHA256 XML signatures, encrypted identity storage, eID workflows, and native dependency fixes
• Built external cash machine integrations for POS workflows through a unified abstraction over REST/JSON and SOAP/XML APIs
• Worked across backend, frontend, infrastructure, AI tooling, analytics, server migrations, deployment reliability, customer-specific product development, and operational planning

Bootstrapped, team of 3 engineers · forest.host · Product later acquired

• Co-built a developer-focused cloud hosting and deployment platform from concept to production
• Built zero-config Git-based deployment workflows, Git integrations for GitHub/GitLab/Bitbucket, public API functionality, and deployment automation
• Owned full-stack and infrastructure work across React, Node.js, CI/CD, LXC/LXD containers, deployment orchestration, backups, rollback systems, SSL, and environment management
• Scaled the platform to 1,000+ users and 10,000+ deployments while balancing architecture, infrastructure cost, reliability, and maintainability
• Product was later acquired