Candidate summary
Halil İbrahim Kurnaz
AI Software Engineer · Istanbul, Turkey
I'm an AI Software Engineer specializing in agentic AI ecosystems, backend architecture, and production-ready GenAI platforms. I architect scalable RAG pipelines, multi-agent orchestrations, and robust tool-calling workflows that turn raw LLM capabilities into real, enterprise-grade products.
Experience
- Product Management Bootcamp Participant · TurkNetSeptember 2026 to November 2026
- Backend AI Engineer Intern · FlyRank AIJune 2026 to September 2026
- AI Engineer · AI Business SchoolNovember 2025 to February 2026
- Business Analyst & Conversational AI Designer · CBOTOctober 2024 to January 2025
- Intern, Flo Gen Digital Internship Program · FLO GroupAugust to September 2026
- Intern, QNB 101 Online Internship Program · QNB TürkiyeAugust to September 2026
Flagship projects
- FlowOne: AI Agent Orchestration Platform · Lead Architect & DeveloperLead architect of an AI-native SaaS platform with Agent Studio, Workflow Studio, and Voice AI.https://useflowone.com/
- PinMind: Multi-Agent Orchestration Framework · Creator & MaintainerA LangGraph-based framework for context-aware agent routing, planning and task orchestration, with reusable workflows, shared memory and an extensible tool-integration layer for autonomous, multi-step execution.https://github.com/hhalilikurnaz/pinmind-web
- Drone Fire Watch: Aerial Fire & Smoke Detection · Graduation ProjectGraduation project: a YOLOv8-L detector trained on the 21K-image D-Fire dataset (precision, recall and mAP50 ≈ 0.80), served by FastAPI to a React operations dashboard and a Streamlit field reporter.https://github.com/hhalilikurnaz
Skills
- AI & Agentic Systems
- LLMs & Agentic AI, RAG Architectures, LangGraph, Tool-Calling, Prompt Engineering, Deep Learning, Computer Vision, ElevenLabs & Twilio
- Backend
- Python, Node.js, FastAPI, SQL, Java, C/C++
- Frontend
- React, TypeScript, JavaScript, Flutter, Next.js
- Infrastructure
- Docker, PostgreSQL, MongoDB, Alembic, Cloudflare R2, CI/CD, Vercel
Education
B.Sc. in Software EngineeringIstanbul Aydın University · Graduated 202614 certificates (AI, cloud, data), verifiable on the portfolio
Suggested interview questions
- FlowOne promises a provider outage never drops a request. Walk us through what happens to one request when the model API returns 503s.
- How is retrieval isolated per tenant in FlowOne's pgvector layer, and how would you prove it in a test?
- Drone Fire Watch reached mAP50 ≈ 0.80. What limited it, and which false positives were hardest to remove?
- Your job system uses lock-based idempotency and 3 retries with backoff. What happens if a worker dies while holding the lock?
- This portfolio splits job-fit analysis into four agents with an offline fallback. Why agents instead of one prompt, and what does it cost?