Ollama: Self-Hosted Large Language Models Replacing OpenAI and Claude APIs Training Course
Ollama is an open-source tool for running large language models locally on consumer and enterprise hardware. It abstracts model quantization, GPU allocation, and API serving into a single command-line interface, enabling organizations to self-host LLMs like Llama, Mistral, and Qwen without sending prompts or data to OpenAI, Anthropic, or Google.
This instructor-led, live training (online or onsite) is aimed at intermediate AI engineers and platform operators who wish to use Ollama to replace cloud LLM APIs with self-hosted, sovereign language model inference.
By the end of this training, participants will be able to:
- Install Ollama on Linux, macOS, and Windows with GPU support.
- Pull, quantize, and serve models from the Ollama registry and HuggingFace.
- Build custom Modelfiles with system prompts and parameter tuning.
- Integrate local LLMs with applications via the OpenAI-compatible API.
- Optimize inference performance for CPU-only and multi-GPU setups.
Format of the Course
- Interactive lecture and discussion.
- Lots of exercises and practice.
- Hands-on implementation in a live-lab environment.
Course Customization Options
- To request a customized training for this course, please contact us to arrange.
Course Outline
AI Sovereignty and LLM Local Deployment
- Risks of cloud LLMs: data retention, training on inputs, foreign jurisdiction.
- Ollama architecture: model server, registry, and OpenAI-compatible API.
- Comparison with vLLM, llama.cpp, and Text Generation Inference.
- Model licensing: Llama, Mistral, Qwen, and Gemma terms.
Installation and Hardware Setup
- Installing Ollama on Linux with CUDA and ROCm support.
- CPU-only fallback and AVX/AVX2 optimization.
- Docker deployment and persistent volume mapping.
- Multi-GPU setup and VRAM allocation strategies.
Model Management
- Pulling models from the Ollama registry: ollama pull llama3.
- Importing GGUF models from HuggingFace and TheBloke.
- Quantization levels: Q4_K_M, Q5_K_M, Q8_0 tradeoffs.
- Model switching and concurrent model loading limits.
Custom Modelfiles
- Writing Modelfile syntax: FROM, PARAMETER, SYSTEM, TEMPLATE.
- Temperature, top_p, and repeat_penalty tuning.
- System prompt engineering for role-specific behavior.
- Creating and publishing custom models to local registry.
API Integration
- OpenAI-compatible /v1/chat/completions endpoint.
- Streaming responses and JSON mode.
- Integrating with LangChain, LlamaIndex, and custom apps.
- Authentication and rate limiting with reverse proxy.
Performance Optimization
- Context window sizing and KV cache management.
- Batch inference and parallel request handling.
- CPU thread allocation and NUMA awareness.
- Monitoring GPU utilization and memory pressure.
Security and Compliance
- Network isolation for model serving endpoints.
- Input filtering and output moderation pipelines.
- Audit logging of prompts and completions.
- Model provenance and hash verification.
Requirements
- Intermediate Linux and container administration.
- Understanding of machine learning and transformer models at high level.
- Familiarity with REST APIs and JSON.
Audience
- AI engineers and developers replacing cloud LLM APIs.
- Organizations with data sensitivity preventing cloud model usage.
- Government and defense teams requiring air-gapped language models.
Need help picking the right course?
southafrica@nobleprog.co.za or +27 (0)10 005 5793
Ollama: Self-Hosted Large Language Models Replacing OpenAI and Claude APIs Training Course - Enquiry
Related Courses
Advanced Ollama Model Debugging & Evaluation
35 HoursAdvanced Ollama Model Debugging and Evaluation is a comprehensive course designed to diagnose, test, and measure the behaviour of models during local or private Ollama deployments.
This instructor-led, live training (available online or on-site) targets advanced-level AI engineers, ML Ops professionals, and QA practitioners who wish to guarantee the reliability, fidelity, and operational readiness of Ollama-based models in production environments.
By the conclusion of this training, participants will be able to:
- Conduct systematic debugging of Ollama-hosted models and reliably reproduce failure modes.
- Design and execute robust evaluation pipelines incorporating quantitative and qualitative metrics.
- Implement observability measures (logs, traces, metrics) to monitor model health and detect drift.
- Automate testing, validation, and regression checks integrated into CI/CD pipelines.
Course Format
- Interactive lectures and discussions.
- Hands-on labs and debugging exercises utilising Ollama deployments.
- Case studies, group troubleshooting sessions, and automation workshops.
Course Customisation Options
- To request customised training for this course, please contact us to make arrangements.
Building Private AI Workflows with Ollama
14 HoursThis instructor-led, live training in Botswana (online or on-site) is tailored for senior professionals seeking to implement secure and efficient AI-driven workflows using Ollama.
Upon completion of this training, participants will be able to:
- Deploy and configure Ollama for private AI processing.
- Integrate AI models into secure enterprise workflows.
- Optimise AI performance while safeguarding data privacy.
- Automate business processes using on-premise AI capabilities.
- Ensure compliance with enterprise security and governance policies.
Deploying and Optimizing LLMs with Ollama
14 HoursThis instructor-led, live training in Botswana (online or onsite) targets intermediate-level professionals seeking to deploy, optimise, and integrate LLMs using Ollama.
Upon completion of this training, participants will be able to:
- Set up and deploy LLMs using Ollama.
- Optimise AI models for enhanced performance and efficiency.
- Leverage GPU acceleration to improve inference speeds.
- Integrate Ollama into workflows and applications.
- Monitor and maintain AI model performance over time.
EXO: End-to-End Local AI Cluster Deployment
21 HoursThis instructor-led, live training in Botswana (online or onsite) is aimed at system administrators and DevOps engineers who wish to deploy, configure, and manage EXO clusters for private LLM inference across multiple Apple Silicon or Linux nodes.
EXO for DevOps: Building Private AI Infrastructure
21 HoursThis instructor-led, live training in Botswana (online or onsite) is aimed at DevOps engineers and infrastructure architects who wish to automate the provisioning, monitoring, and lifecycle management of private AI clusters built on EXO.
EXO Security and Governance: Offline Model Management
14 HoursThis instructor-led, live training in Botswana (online or onsite) is aimed at security engineers and compliance officers who wish to harden EXO deployments, control model access, and govern AI workloads running entirely on-premise.
Fine-Tuning and Customizing AI Models on Ollama
14 HoursThis instructor-led, live training in Botswana (online or onsite) is aimed at advanced-level professionals who wish to fine-tune and customize AI models on Ollama for enhanced performance and domain-specific applications.
By the end of this training, participants will be able to:
- Set up an efficient environment for fine-tuning AI models on Ollama.
- Prepare datasets for supervised fine-tuning and reinforcement learning.
- Optimize AI models for performance, accuracy, and efficiency.
- Deploy customized models in production environments.
- Evaluate model improvements and ensure robustness.
Secure Local Agentic AI: On-Prem Ollama Development for Regulated Industries
21 HoursThis instructor-led live training in Botswana (online or onsite) is targeted at developers and technical teams who wish to utilise Ollama and open models to build and run private agentic AI solutions on internal infrastructure.
By the end of this training, participants will be able to install and configure Ollama, evaluate and operate open models locally, create simple agentic and retrieval-based workflows, and apply security and governance controls suitable for regulated environments.
Multimodal Applications with Ollama
21 HoursOllama is a platform that allows you to run and fine-tune large language and multimodal models on your own machine.
This instructor-led, live training (available online or onsite) is designed for advanced-level ML engineers, AI researchers, and product developers who want to build and deploy multimodal applications using Ollama.
By the end of this training, participants will be able to:
- Set up and run multimodal models with Ollama.
- Integrate text, image, and audio inputs for real-world applications.
- Build document understanding and visual QA systems.
- Develop multimodal agents capable of reasoning across modalities.
Format of the Course
- Interactive lecture and discussion.
- Hands-on practice with real multimodal datasets.
- Live-lab implementation of multimodal pipelines using Ollama.
Course Customization Options
- To request a customized training for this course, please contact us to arrange.
Getting Started with Ollama: Running Local AI Models
7 HoursThis instructor-led, live training in Botswana (online or onsite) is aimed at beginner-level professionals who wish to install, configure, and use Ollama for running AI models on their local machines.
By the end of this training, participants will be able to:
- Understand the fundamentals of Ollama and its capabilities.
- Set up Ollama for running local AI models.
- Deploy and interact with LLMs using Ollama.
- Optimize performance and resource usage for AI workloads.
- Explore use cases for local AI deployment in various industries.
Ollama & Data Privacy: Secure Deployment Patterns
14 HoursOllama serves as a platform enabling the local execution of large language and multimodal models, while facilitating robust secure deployment strategies.
This instructor-led live training, available both online and onsite, is designed for intermediate-level professionals aiming to deploy Ollama with stringent data privacy and regulatory compliance measures.
Upon completion of this training, participants will be capable of:
- Securely deploying Ollama within containerized and on-premises environments.
- Applying differential privacy techniques to protect sensitive data.
- Implementing secure logging, monitoring, and auditing practices.
- Enforcing data access controls that align with compliance requirements.
Format of the Course
- Interactive lectures and discussions.
- Hands-on labs focusing on secure deployment patterns.
- Compliance-oriented case studies and practical exercises.
Course Customization Options
- To request customized training for this course, please contact us to make arrangements.
Ollama Applications in Finance
14 HoursOllama is a lightweight platform designed for running large language models locally.
This instructor-led training session, available either online or onsite, is tailored for intermediate-level finance professionals and IT staff who wish to implement, customise, and operationalise Ollama-based AI solutions within financial settings.
Upon completing this training, participants will acquire the skills required to:
- Deploy and configure Ollama to ensure secure use in financial operations.
- Integrate local LLMs into analytical and reporting workflows.
- Adapt models to finance-specific terminology and tasks.
- Apply security, privacy, and compliance best practices.
Course Format
- Interactive lectures and discussions.
- Hands-on exercises using financial data.
- Live-lab implementation of finance-focused scenarios.
Customisation Options
- To request a customised training programme for this course, please contact us to make arrangements.
Ollama Applications in Healthcare
14 HoursOllama is a slim platform designed for running large language models on your own devices.
This instructor-led, live training (available online or on-site) targets intermediate healthcare professionals and IT teams who want to deploy, tailor, and run Ollama-based AI solutions in clinical and administrative settings.
After finishing this training, participants will be able to:
- Set up Ollama for secure use in healthcare environments.
- Integrate local LLMs into clinical workflows and admin processes.
- Tailor models for healthcare-specific language and tasks.
- Apply best practices for privacy, security, and regulatory compliance.
Course Format
- Interactive lectures and discussions.
- Hands-on demonstrations and guided exercises.
- Practical implementation in a sandboxed healthcare simulation environment.
Customisation Options
- To request a customised training for this course, please contact us to arrange.
Ollama for Responsible AI and Governance
14 HoursOllama serves as a platform for executing large language and multimodal models locally, thereby supporting governance and responsible AI practices.
This instructor-led, live training, available online or on-site, is designed for intermediate to advanced professionals who wish to implement fairness, transparency, and accountability in applications powered by Ollama.
Upon completing this training, participants will be able to:
- Apply responsible AI principles within Ollama deployments.
- Implement strategies for content filtering and bias mitigation.
- Design governance workflows to ensure AI alignment and auditability.
- Establish monitoring and reporting frameworks to ensure compliance.
Course Format
- Interactive lectures and discussions.
- Hands-on labs focused on designing governance workflows.
- Case studies and exercises centred on compliance.
Course Customisation Options
- For a customized training version of this course, please contact us to make arrangements.
Sovereign AI for Regulated Organizations: Controlling Data, Models and Inference Environments
7 HoursThis instructor-led, live training in Botswana (online or onsite) is aimed at intermediate-level IT leaders, compliance professionals, security teams, and enterprise architects who wish to use sovereign AI principles and governance practices to design AI environments that protect sensitive data, support localization requirements, and reduce vendor lock-in.
By the end of this training, participants will be able to: explain sovereign AI concepts, evaluate hosting and governance options, define controls for prompts and logs, and create a practical adoption roadmap.
