CS Graduate · Backend Engineer · Open Source Builder
Building practical systems, one problem at a time
C++ Go Python
IoT MQTT Distributed Systems
Production-ready backend systems that run on real constraints.
No hype. Just working code.
Current driver:
→ agrisense – open-source IoT monitoring platform
- 2,300+ req/s, <6ms P95 latency
- Device management, rule-based alerting, full Docker deployment
- Runs on a $10 VPS – built for small farms and edge environments
I'm a CS graduate passionate about building efficient, scalable backend systems. I love contributing to open source and exploring low-level programming challenges. Currently seeking full-time opportunities in backend development and systems design.
- 🌱 Always diving deeper into systems programming and architecture
- 🤝 Happy to collaborate on interesting projects—reach out any time
- 🎯 Open to work — actively seeking backend/systems development, AI, or anything that sparks curiosity
| Project | Description | Stack |
|---|---|---|
| agrisense | Production IoT platform – data ingestion, alerts, dashboard | Go, EMQX, PostgreSQL, InfluxDB, Redis, Docker |
| high-performance-server | HTTP server from scratch (Epoll ET + thread pool + timer) | C++, Linux, Epoll |
| Student Management System | Cross-platform desktop app with multi-language, export | C++, Qt6, SQLite |
More on repositories.
I'm curious about making IoT systems smarter without expensive hardware:
- Session memory for agents: How to retain context without resending full history? → experimenting with RAG + vector DB on CPU.
- Low-cost anomaly detection: Can a tiny LLM (3B) on a $10 VPS suggest useful alert thresholds? → learning Ollama + Go embedding.
These are learning explorations. I'll share what works (and what fails).
- Solve one concrete problem at a time
- Fast iteration, then public documentation
- Learn only what the current problem requires
I don't post much, but I reply to real issues and genuine collaboration.
- Email: atp19693@gmail.com
- GitHub: savvyinsight
- LinkedIn: Abudukaiyumu Yasen
"Build it. Run it. Document it. Then move to the next problem."