Arif Dogan — backend & AI engineer.

Backend & AI engineer. I ship production systems for teams that need to move fast — from LLM features to the correctness-critical logic underneath.

★ Open source

5 tools.
1.2k stars on GitHub.

1.2kgithub · 2024 — present

MVP buildIdea and budget in hand — I turn it into a deployed, production-ready product.4–8 wks · fixed price

AI integrationLLM orchestration, agents, RAG, and evaluation added to your product. Production-grade, not a wrapper.2–4 wks · milestone

Fractional CTO / tech leadArchitecture, hiring, code review, vendor calls. Senior technical leadership, part-time.ongoing · weekly

GoVisualZero-config HTTP request visualizer for Go — debug complex services without leaving the terminal682 ★

FastAPI RadarReal-time monitoring and visualization for FastAPI services in production400 ★

LLMDogContext preparation toolkit that keeps LLM responses reliable for complex inputs81 ★

K9SightKeyboard-driven TUI for debugging Kubernetes workloads — pods, logs, exec, deployments65 ★

GIQAI-enhanced Git workflow assistant that automates code review preparation15 ★

Structuring a managed search engine in RustControl plane, data plane, durable storage: one search engine split across three systemsarticle

Shipping LLM features that survive productionWhy most LLM features break in the real world — and the evaluation layer that keeps them honestarticle

Turning a regulation into testable codeThe method I use to take a statute from primary source to audit-ready detectorarticle

Building a GDPR-compliant backendAn engineer's checklist — the architecture decisions that actually carry legal weightarticle

The EU AI Act for software engineersWhat you actually have to build — obligations translated into system requirementsarticle

The Context TaxWhy AI-assisted coding fails without flowessay

I write about shipping production systems and AI you can trust — the same lens I bring to client work.

MCP in the Wild: Cross-Domain Knowledge Discovery through Multi-Server OrchestrationAn empirical study of multi-server MCP orchestration — 17 real tool calls across 6 servers, a 7-model benchmark, and 5 composition patterns. Preprint.2026

Discovery of TIC 63038821: A New UV Ceti / BY Draconis Variable M-Dwarf Binary from TESS Sector 1Discovered a previously unreported variable star using autoencoder anomaly detection on 15,665 TESS light curves. Flare events, rotational modulation, chromospheric activity confirmed.2026