# Dawit Berhanu — AI Product Engineer
# LLM-optimized summary · https://davesovm.xyz/
# Full version: https://davesovm.xyz/llms-full.txt
# Last updated: 2025-06-14
# Standard: https://llmstxt.org

## Identity
Name: Dawit Berhanu
Handle: davesovm
Title: AI Product Engineer
Location: Addis Ababa, Ethiopia (UTC+3)
Website: https://davesovm.xyz/
Email: dawitsovm@gmail.com

## About
I build developer infrastructure to replace friction with flow. My work focuses
on systems that act as an extension of your own thought process. I'm currently
building TEBIBX to turn repository complexity into actionable intelligence.

## Company
Name: SOVM'S
Type: AI Infrastructure Company
Mission: Make every codebase instantly understandable.
Focus: Developer intelligence infrastructure — turning repositories into
structured, searchable, and explainable systems.

## Flagship Product: TEBIBX
URL: https://tebibx.dev
Type: AI-Native Repository Intelligence Platform
Description: Transforms any GitHub repository into an intelligent system that
can be searched, understood, and navigated using natural language.
Outcomes: 500+ queries answered across beta repos · supports 20+ repositories
Capabilities:
- Natural language questions about any codebase
- Instant architecture and logic summaries
- Dependency and structure visualization
- Understanding large systems in seconds
Ecosystem:
- Dashboard (web) — https://tebibx.dev · Demo: https://www.youtube.com/watch?v=g44L07PCrjM
- Browser Extension — https://tebibx.dev · Demo: https://www.youtube.com/watch?v=SsqaUi1xlyc
- CLI (terminal) — https://tebibx.dev
- VS Code Extension — https://tebibx.dev

## Secondary Product: WISDOM PASS
URL: https://wizdompass.com
Type: Knowledge Infrastructure Platform
Description: Structured knowledge system for engineering teams. Captures,
organizes, and surfaces institutional knowledge to reduce knowledge loss and
accelerate onboarding.
Live analytics report: https://lookerstudio.google.com/reporting/63a64a51-797b-41d7-9b0b-c92f418765aa

## Open Source: TBXpost
URL: https://github.com/davesovm/TBXpost
Type: Personal content OS for developers
Description: Logs dev work, curates tech news via Groq AI, generates posts in
your voice, and auto-posts to Telegram twice daily. Runs entirely on GitHub
Actions — zero cloud cost.
Telegram channel: https://t.me/sovm00

## Case Study: TEBIBX Architecture
URL: https://davesovm.xyz/case-study.html
Title: When your server has 512MB and your users have entire codebases
Summary: How a 512MB RAM constraint on a Render instance led to moving the
entire RAG ingestion pipeline to the browser (Web Worker + IndexedDB), keeping
the server stateless during ingestion while maintaining a 6-stage hybrid
retrieval pipeline (Qdrant semantic + SQLite BM25 + call graph + map/reduce
with Gemini Flash and Groq Llama).

## Skills
- AI systems and LLM tooling
- SaaS product engineering
- Backend infrastructure (FastAPI, Python, Rust)
- Developer tooling (CLI, VS Code extensions, browser extensions)
- React, TypeScript, Supabase, Redis, Qdrant
- RAG, LLM integration, vector databases
- GitHub Actions automation and CI/CD pipelines
- Telegram Bot API integration

## Social & Content Channels
LinkedIn: https://www.linkedin.com/in/dawit-berhanu/
X/Twitter: https://x.com/DawitSovm
Dev.to: https://dev.to/tebibx
Telegram channel: https://t.me/sovm00
Discord: https://discord.gg/rM6dVvEdM