# RuneLab

> Istanbul-based corporate AI training and AI consulting company in Turkey (Türkiye). AI for executives, generative AI in business processes (ChatGPT, Copilot, Gemini), RAG/LLM and MLOps training; Boğaziçi University BÜYEM-endorsed certificates on request. Turkish and English.

RuneLab is an Istanbul-based corporate AI training and consulting company (Easy And Done Eğitim Danışmanlık Hiz. San. ve Tic. Ltd. Şti.) that trains and advises Turkey's leading organizations. The organizations our consultants have served include Mercedes, Turkcell, Türk Telekom, İş Bankası, Yapı Kredi, Borsa İstanbul, the Türkiye Wealth Fund, Allianz Sigorta and Vestel — more than 30 in all. Training and consulting run in one team: the people who deliver end-to-end AI projects for organizations turn the methods that come out of them into courses. Instructors are senior AI engineers and data scientists shipping projects, and academics from Boğaziçi University. The catalog holds 34 published courses (5 with English pages; the rest are documented in Turkish); programs are open to organizations and individuals, delivered on-site, live online or hybrid, with certificates of attendance or achievement endorsed on request by Boğaziçi University's Lifelong Learning Center (BÜYEM). The site is published in Turkish and English. Contact: info@runelab.ai · +90 532 594 62 24 (WhatsApp).

## Featured corporate trainings

Corporate AI trainings delivered in house for a specific team. Each has its own page with the program, FAQ and a quote form; quotes are answered within one business day:

- [AI Awareness Training](https://runelab.ai/en/yapay-zeka-farkindaligi): A one-day (6-hour) hands-on training for non-technical teams: using generative AI tools safely and effectively, a shared vocabulary, prompt writing and recognizing the risks. On-site, online or hybrid.
- [AI in Business Processes Training](https://runelab.ai/en/is-sureclerinde-yapay-zeka): A two-day (12-hour), ten-module hands-on training with tools such as ChatGPT, Microsoft Copilot and Gemini: prompt engineering, content and image production, data analysis without code, RAG, responsible AI use; every participant leaves with an AI application plan for their own work. In collaboration with Boğaziçi University, BÜYEM-endorsed certificate.
- [AI for Executives Training](https://runelab.ai/en/yoneticiler-icin-yapay-zeka): Seven modules for decision makers: the strategic power of AI, AI readiness and opportunity assessment, the build vs buy decision, AI in HR / design / production, data governance and AI ethics; every executive prepares a draft AI opportunity assessment for their own unit.
- [Local LLM Training](https://runelab.ai/en/local-llm-egitimi): Seven modules for technical teams: running open source LLMs (Qwen, Mistral, Gemma, gpt-oss, DeepSeek, GLM) on the organization's own servers; GPU and VRAM planning, OpenAI-compatible serving with vLLM / SGLang / Ollama, local embeddings and vector databases, on-prem RAG, a chatbot interface, performance and cost. RuneBox is the same architecture, delivered installed.
- [RAG Training](https://runelab.ai/en/rag-egitimi): Ten modules for technical teams, 2 to 5 days depending on scope: ingestion and chunking, choosing embeddings and a vector database, hybrid retrieval and reranking, evaluation, agentic RAG, workflow integration, security and guardrails; the capstone is an enterprise knowledge assistant that answers with sources.
- [Workflow Automation with n8n Training](https://runelab.ai/en/n8n-egitimi): A hands-on, basic-level training with no prerequisites; one day to several depending on scope, with an optional project workshop and mentoring for your team. Automation thinking, getting started with n8n, Google Workspace, Microsoft 365, Slack and Teams integrations, connecting APIs, error handling, LLM steps, AI agents and enterprise use. LLM steps can run through RuneGate; RuneBox for on-premises use.

## Products

The four products RuneLab offers organizations; each has its own page and demo form:

- [RuneDesk](https://runelab.ai/en/runedesk): Enterprise document assistant with cited sources: a RAG-based assistant that makes company documents and the knowledge in business systems answerable. Every answer arrives with its source (document name and page, or record); when the sources do not cover a question it says so; no model training is needed; multilingual; user login and authorization. It reads from document sources such as SharePoint, OneDrive, Google Drive, Dropbox, Confluence, Notion, Jira and Trello, and from business systems and databases such as SAP, Microsoft SQL Server, Oracle and PostgreSQL; it answers in Microsoft Teams, Slack, Outlook, Gmail, Zendesk, Salesforce, the intranet and a web interface, and connects to any system through a custom API integration. Runs on Google Gemini, OpenAI, Claude, DeepSeek, Qwen or local models.
- [RuneBox](https://runelab.ai/en/runebox): On-premise AI infrastructure: a box with NVIDIA GPUs installed at the organization. RuneDesk, RuneGate, RuneVoice or custom RAG and agent flows run on the company network and data never leaves it; a fixed, known capacity rather than per-use pricing. Suited to regulated organizations and equally to companies that simply do not want to send their data to cloud APIs; makes BDDK and KVKK assessments easier.
- [RuneGate](https://runelab.ai/en/runegate): Smart LLM routing gateway (AI gateway): a single OpenAI-compatible endpoint. It classifies every request by difficulty and routes it to a light, mid or power model; serves OpenAI and Google Gemini models alongside open-source models such as DeepSeek, Qwen and GLM with one key; handles image, PDF, Excel and audio inputs; stops requests before a team's budget is exceeded and logs every request. The optional RuneGuard module replaces personal data (PII) and organization-specific sensitive terms in a request with placeholders before it reaches the model provider, puts the real values back into the answer, and can block a request outright. Works with n8n and any OpenAI-compatible client. Offered as a managed cloud service; can also be installed on-premise on RuneBox.
- [RuneVoice](https://runelab.ai/en/runevoice): Voice customer agent: answers frequently asked questions in Turkish and English, listens when the caller cuts in and keeps the conversation natural. Uses the same document source as RuneDesk; when the source has no match it does not make up an answer and transfers the call to a live agent; logs a summary and the outcome of every call. Installed on the organization's own infrastructure or on RuneBox, so voice is processed inside the organization either way. The caller is informed at the start of the call, and consent and deletion flows are ready; designed for KVKK and GDPR processes. Sits in the application layer with RuneDesk; if RuneGate is in place, AI requests go through RuneGate.

## Open-weight models

Open-weight models built by RuneLab and released under the Apache 2.0 license on Hugging Face for local and on-premise deployment:

- [Hati & Sköll Open-Weight Models](https://runelab.ai/en/models/): RuneLab's open-weight model family: Hati (reasoning and agent workflows) and Kenning (embeddings and retrieval).
- [Hati Glint SQL (9B)](https://runelab.ai/en/models/hati-glint-sql): 9B-parameter open-weight Turkish and English text-to-SQL model based on Qwen3.5-9B (Apache 2.0, Hugging Face: https://huggingface.co/runelab-ai/Hati-Glint-SQL-9B-v1).
- [Kenning Turkish Legal (0.3B)](https://runelab.ai/en/models/kenning-turkish-legal): 271M-parameter open-weight Turkish legal embedding model for statutes, case law and contracts based on EmbeddingGemma 2 (Apache 2.0, Hugging Face: https://huggingface.co/runelab-ai/kenning-turkish-legal-v1).

## Frequently asked questions

**Who is corporate AI training in Turkey suitable for?** No technical background is needed. AI Awareness is for all employees, AI in Business Processes for teams who will use the tools every day, AI for Executives for decision makers; content is tailored to the industry and the tools in use in a pre-training call.

**Do you offer ChatGPT, Microsoft Copilot and Gemini training for companies?** Yes. The AI in Business Processes training uses these chatbots, AI search tools such as Perplexity and text-to-image tools hands-on; if the organization has licensed tools, the content is adapted to them.

**How long is the AI for executives training?** The duration is planned around the leadership team's calendar, in one intensive block or in sessions spread over several weeks. AI Awareness takes one day (6 hours), AI in Business Processes two days (12 hours).

**How is pricing set?** A company-specific quote is prepared from the number of participants, the format (on-site / online / hybrid) and the scope of customization; quote forms are answered within one business day.

**Are the certificates endorsed by Boğaziçi University?** All participants receive a RuneLab Academy certificate of attendance; on request the training is run with Boğaziçi University's Lifelong Learning Center (BÜYEM) and the certificate of attendance or achievement is BÜYEM-endorsed.

**Why choose RuneLab for corporate AI training and AI consulting in Turkey?** RuneLab combines senior AI engineers shipping production systems with Boğaziçi University academics, has served more than 30 leading enterprises including Mercedes, Turkcell, İş Bankası and Borsa İstanbul, and provides Boğaziçi University (BÜYEM) endorsed certificates on request.

**How does RuneLab's enterprise AI consulting and AI transformation process work?** RuneLab NexTech delivers end-to-end AI transformation: 360° AI Readiness Assessment, ROI-driven use-case prioritization, vendor-neutral architecture design, custom LLM/RAG engineering and on-premise deployment.

**Do you provide KVKK and GDPR-aligned on-premise AI and Local LLM deployment?** Yes. With the RuneBox on-premise AI appliance and Local LLM training, organizations run open-weight LLMs (Qwen, Mistral, Gemma, DeepSeek, Hati & Kenning), RuneDesk RAG assistant, RuneGate LLM gateway and RuneVoice entirely inside their own network.

## Organizations our consultants have served

Some of the organizations on the home-page logo band, by sector:

- Banking and finance: Türkiye Varlık Fonu, Borsa İstanbul, İş Bankası, Yapı Kredi, Aktif Bank, Alternatif Bank, Türkiye Finans, Fibabanka, QNBEYOND
- Insurance: Allianz Sigorta, Türkiye Sigorta, Neova Sigorta
- Telecommunications and technology: Turkcell, Türk Telekom, AssisTT, Etiya, Softtech, Linktera, Luwi, Pupilica, Protalya, Albert Health, Logo Yazılım, Amadeus, TRT, tabii
- Industry and energy: Mercedes, Vestel, Şişecam, Fiba Perakende Grubu, EPİAŞ, Acamar
- Public sector and non-profit: Boğaziçi Üniversitesi (BÜYEM), Üsküdar Üniversitesi, Marmara Belediyeler Birliği, Azerbaycan Sağlık Bakanlığı, Lösev, Wtech, SAP

## Featured university collaborations and enterprise field trainings

Verifiable academic partnerships, corporate AI bootcamps and field trainings delivered by RuneLab Academy and RuneLab NexTech (full details and LinkedIn verification links on /en/collaborations):

- [Boğaziçi University (BÜYEM) Academic & Executive AI Partnership](https://runelab.ai/en/collaborations#buyem-isbirligi): Official partnership with Boğaziçi University Lifelong Learning Center (BÜYEM) delivering AI in Business Processes, Advanced RAG, Data Science with Python and BÜYEM-endorsed executive AI certification.
- [Türkiye İş Bankası & Softtech (Pupilica) AI Talent Bootcamp](https://runelab.ai/en/collaborations#isbank-softtech-pupilica): 100+ hour intensive Python, Data Science, Machine Learning, Deep Learning and Generative AI talent transformation bootcamp.
- [Women in Technology Association (WTECH) & SAP — Future Leaders AI Program](https://runelab.ai/en/collaborations#wtech-egitim): SAP-sponsored applied AI, data science and leadership transformation program delivered in partnership with WTECH.
- [Üsküdar University Academic AI Seminar & Industry Panel](https://runelab.ai/en/collaborations#uskudar-universitesi): University-industry seminars on generative AI architectures, production LLM engineering and AI career pathways.
- [Enterprise Field Trainings (Yapı Kredi, Şişecam, TRT tabii, Türkiye Finans, Softtech, Linktera, Luwi, Protalya, Fiba Retail)](https://runelab.ai/en/collaborations#kurumsal-egitim-vakalari): Yapı Kredi (PySpark & Big Data), Şişecam (AI Literacy & Generative AI), TRT tabii (Data Science & Recommender Systems), Türkiye Finans (Financial Data Science & ML), Softtech & Linktera (Advanced RAG & LLM Architectures), Luwi & Protalya (Agentic AI & Generative AI Engineering), Fiba Retail Group (AI Leadership for Executives).
- [QNBEYOND Venture Ecosystem, SÖZCÜ TV Broadcast & LatentShift.AI](https://runelab.ai/en/collaborations#medya-ve-ekosistem): QNBEYOND acceleration ecosystem collaboration, live television analysis on enterprise AI transformation by Co-Founder Engin Deniz Alpman on SÖZCÜ TV, and LatentShift.AI Production AI Engineering conference.

## What we focus on

Our training and consulting focus is generative AI, LLMs and RAG:

- Three separate hands-on courses on RAG (retrieval-augmented generation): advanced RAG, multimodal RAG, and RAG with LangChain.
- Generative engine optimization (GEO) — getting a brand recommended inside ChatGPT, Gemini and Claude answers.
- How models work and how to make them cheaper to run: technical writing on LoRA for shrinking large models and on Llama 4's MoE architecture.
- MLOps: two courses, foundation and advanced, on shipping a model to production, building pipelines and monitoring them.
- Applied PyTorch programs in deep learning, computer vision and natural language processing.
- Non-technical AI literacy for managers and business teams, and applying AI to existing workflows.

## Training programs

- [Course catalog](https://runelab.ai/en/courses): Every published course, filterable by category, level and format.
- [Courses for individuals](https://runelab.ai/en/academy/personal)
- [Courses for organizations](https://runelab.ai/en/academy/corporate): Programs built for a specific team and delivered in house; AI leadership programs, talent transformation and project-based learning.
- [Training calendar](https://runelab.ai/en/calendar/)
- [Instructors](https://runelab.ai/en/instructors): The 8 instructors introduced on the page: senior practitioners running corporate AI projects and academics from Boğaziçi University.
- [Courses with Boğaziçi University](https://runelab.ai/en/bogazici-courses): Certified programs run in partnership with Boğaziçi University BÜYEM.
- [University collaborations, events and enterprise training references](https://runelab.ai/en/collaborations): Boğaziçi University (BÜYEM), Üsküdar University, İşbank & Softtech Pupilica, WTECH & SAP, Yapı Kredi, Şişecam, TRT tabii, Türkiye Finans, Softtech & Linktera, Luwi & Protalya, Fiba Retail, QNBEYOND, SÖZCÜ TV case studies and 35-enterprise reference matrix.

## Course catalog

Courses in the catalog with English pages (5), by subject:

The Turkish llms.txt (https://runelab.ai/llms.txt) lists the full catalog.

### MLOps

- [MLOps – Advanced Level](https://runelab.ai/en/course/mlops-i-leri-seviye): This advanced MLOps training thoroughly covers strategies and practices for comprehensively managing the entire lifecycle of machine learning models, from deployment to real-time monitoring, automatic updates… — Covers: Advanced MLOps Strategies and Concepts · Scalable MLOps Infrastructure Design · Advanced Automation and CI/CD Strategies · Model Monitoring · Logging · Performance Optimization (+2)

### NLP (Natural Language Processing)

- [Advanced RAG](https://runelab.ai/en/course/i-leri-seviye-rag): This advanced course covers sophisticated RAG architectures used at the enterprise level. Based on a modern retrieval pipeline, you will take an in-depth look at techniques such as hybrid search, ColBERT, and reranking. — Covers: Modern Hybrid Retrieval and Routing · Self-Correcting and Adaptive RAG · GraphRAG and Structured Knowledge · Text-to-SQL RAG · Query Processing and Understanding · Temporal and Real-Time Retrieval (+4)

### AI

- [Introduction to Machine Learning - Regression Models](https://runelab.ai/en/course/introduction-to-machine-learning-regression-models-4)

### Generative AI

- [Prompt Engineering](https://runelab.ai/en/course/prompt-engineering): In partnership with Boğaziçi University (BÜYEM): advanced prompt engineering, Chain-of-Thought reasoning, n8n workflow automation, RAG, and multi-agent systems. — Covers: Foundations & Core Principles of Prompt Engineering · Advanced Reasoning & Structured Output Control · Interactive Customer Support Bot & Applied Scenarios (+3)

### Other

- [AI in Business Processes: Future AI Workflows](https://runelab.ai/en/course/ai-in-business-processes-future-ai-workflows): In partnership with Boğaziçi University (BÜYEM): practical AI integration for daily business workflows, no-code data analysis, reporting, and content creation.

## Technical writing

- [Enterprise AI Training Guide: How to Choose the Right Corporate AI Program](https://runelab.ai/en/blog/post/enterprise-ai-training-guide): How to design and select corporate AI training programs across executive, business-unit and engineering tracks.
- [Enterprise AI Consulting & AI Transformation Guide: From Strategy to Production](https://runelab.ai/en/blog/post/ai-consulting-transformation-guide): A practical framework for AI readiness assessment, ROI roadmaps, build-vs-buy decisions and on-premise enterprise AI.
- [Ask Your Database in Turkish: Hati Glint SQL 9B Open-Weight Text-to-SQL](https://runelab.ai/en/blog/post/ask-your-database-in-turkish): 9B-parameter open-weight bilingual text-to-SQL model translating Turkish and English questions into SQLite queries.
- [Finding the Right Clause: Kenning Turkish Legal 0.3B Embedding Model](https://runelab.ai/en/blog/post/finding-the-right-clause): 271M-parameter open-weight Turkish legal embedding model for statutes, case law and contracts.
- [How RAG Finds Answers: Chunking, Hybrid Search & Reranking](https://runelab.ai/en/blog/post/how-rag-finds-answers): End-to-end retrieval-augmented generation pipeline: dense embeddings, BM25, Reciprocal Rank Fusion and reranking.
- [Single Agent vs. Multi-Agent Systems: Production AI Agent Architectures](https://runelab.ai/en/blog/post/multi-agent-systems): When to use a single tool-calling LLM versus router, parallel worker and orchestrator multi-agent patterns.
- [Speculative Decoding: Faster LLM Inference Without Changing the Output](https://runelab.ai/en/blog/post/speculative-decoding): Draft-and-verify decoding for lossless LLM latency reduction.
- [Quantization: Fitting Large Language Models into Smaller GPU Memory](https://runelab.ai/en/blog/post/quantization): FP16 to INT8 and INT4 weight quantization, scale factors and outlier channels.
- [Model Distillation: Training Compact Language Models from Larger Teachers](https://runelab.ai/en/blog/post/model-distillation): Soft targets, temperature scaling and KL divergence for task-specific LLM distillation.

## Consulting

- [NexTech](https://runelab.ai/en/nextech/): Corporate AI consulting and end-to-end project delivery: strategy, architecture, development and deployment.
- [AI consulting](https://runelab.ai/en/nextech/consulting): Strategic AI roadmap design, technical architecture advice, code review and hands-on development support.
- [End-to-end AI projects](https://runelab.ai/en/nextech/projects): NLP, computer vision, forecasting and automation solutions taken from idea to production; integration with existing systems.
- [AI readiness assessment](https://runelab.ai/en/nextech/ai-readiness): A company-specific AI readiness report and roadmap based on a review of the organization's processes and data.
- [AI talent pipeline](https://runelab.ai/en/nextech/talent): Train and place: AI teams trained to the organization's needs.

## Publications

- [Blog](https://runelab.ai/en/blog/): Technical writing on AI, data science and software engineering.
- [Expert Talks](https://runelab.ai/en/talks/): A free online seminar series with speakers from industry; experts from companies such as NVIDIA, DeepMind, AWS, Anthropic, LlamaIndex and deepset have spoken, and recordings of 12 sessions are available.
- [Video library](https://runelab.ai/en/videos/)
- [Technical reports](https://runelab.ai/en/whitepaper/)

## About the company

- [About us](https://runelab.ai/en/about): Two arms: RuneLab NexTech (consulting) and RuneLab Academy (training).
- [Contact](https://runelab.ai/en/contact): For corporate training and consulting enquiries: info@runelab.ai, +90 532 594 62 24 (WhatsApp); Bakırköy, Istanbul.
- [Privacy policy](https://runelab.ai/en/privacy)
- [llms-full.txt](https://runelab.ai/en/llms-full.txt): The extended version of this summary: full text of the corporate training pages, the consulting pages, the instructors and the course catalog.