
Telegram SME AI Assistant π€
AI-powered Telegram assistant for small businesses: inventory management π¦, invoicing, receipts, and analytics.
I'm a Final Year Computer Science Student from Uganda specialising in modern web development and applied AI. I craft elegant frontends, resilient backends, and integrate LLMs/AI systems that feel seamless, not stitched on. I love performance budgets, design systems, and solving real product problems with code.


AI-powered Telegram assistant for small businesses: inventory management π¦, invoicing, receipts, and analytics.

Cross-platform mobile app for food ordering π, menu browsing, cart π, address management, order tracking, and more across 3 branches.

Deploy, manage, and scale large language models (LLMs) securely within your own infrastructure, no sensitive data leaves your organization.
Deep dives into Tailwind, LangChain, Docker, Express, Hono, and LLM systems β how I use them end to end in real projects.
Writing helps me design clearer systems.
How I use Tailwind to ship fast without creating a mess: constraints, design tokens, and patterns that scale as projects grow.
A practical mental model for LangChain: chains, tools, memory, and how to avoid building fragile LLM spaghetti.
From a single container to multi-service deployments: Compose, networks, volumes, and how I think about environments.
The core Express patterns I actually use in production: routing, middleware, error handling, and project structure.
Why I enjoy Hono for small but sharp APIs: edge runtimes, DX, and when I pick it over Express.
Principles I follow to integrate LLMs into products so they feel native, reliable, and worth keeping.
Thinking in states, not screens: skeletons, optimistic updates, and failure UIs that still feel calm.
How treating university and freelance work like real client projects changed my engineering mindset.