Blog
All articles. All pillars. All authors.
#122 -- Code-Aware Chunking for RAG
How FLIN's chunk_text() function splits documents into embedding-appropriate segments while respecting paragraph boundaries, code blocks, headings, and semantic coherence.
Six Languages, One Platform: i18n for Africa
French, English, Arabic (with RTL), Swahili, Portuguese, Spanish. How we built internationalization for a continent with hundreds of languages.
Organizations: Families, Schools, and Companies on One Platform
Multi-tenant organizations with shared credit pools, join codes, access codes for phoneless students, and role-based membership for families, schools, and companies.
101 AI Advisors: Professional Intelligence for Africa
60+ specialized AI advisors across 14 categories: SYSCOHADA accountants, OHADA lawyers, tax specialists. How we built professional intelligence for Africa.
#001 -- Why We Built a Programming Language From Scratch
Why Juste A. GNIMAVO and AI CTO Claude built FLIN, a programming language replacing 47 technologies, from Abidjan.
RAG Pipeline: Document Search With pgvector and Semantic Chunking
PostgreSQL with pgvector for embeddings, Datalab API for semantic chunking, Mistral Reranker for top-k refinement, and context compression at 150K tokens.
From Abidjan to 250 Million: The Deblo.ai Story
96 web sessions + 20 mobile sessions in 6 weeks. 24+ tables, 100+ endpoints, 24 AI tools, 101+ advisors. The complete story of building Deblo.ai from Abidjan.
#004 -- Five Design Principles That Shape Every Line of FLIN
The five design principles behind FLIN: simple, zero-config, reactive, intent-native, and memory-native.
Background Jobs: When AI Takes 30 Minutes to Think
Queue-bridged architecture with detached asyncio tasks, Redis progress tracking, cooperative cancellation, and stale job cleanup for long-running AI generation.
#019 -- Error Diagnostics That Actually Help
How FLIN's error diagnostics help developers: source locations, colored output, and human-readable messages.
AI Memory and Context Compression
Fire-and-forget conversation summarization, the save_memory tool, and context compression at 150K tokens. How the AI remembers across sessions for $0.00005 per summary.
Writing a Docker Engine Client from Scratch in Rust
Why we wrote a custom Docker Engine API client using hyper and Unix sockets instead of shelling out to the Docker CLI, and the multiplexed stream parsing that made it work.