Udemy Self-Hosting LLMs Securely- Hardening Ollama and vLLM 2026-9

File Name

Size

1 - What You Are Actually Running/1. Course Documents and How to Use Them.mp4

32 MB

1 - What You Are Actually Running/1. D01_Inference_Host_Threat_Model_One_Page_Worksheet_and_Worked_Example.docx

22 kB

1 - What You Are Actually Running/1. D02_Baseline_Exposure_Scoreboard_Measure_Your_Default_Install_Before_You_Change_.docx

21 kB

1 - What You Are Actually Running/1. D03_Ollama_Hardening_Checklist_Host_Service_and_Container.docx

25 kB

1 - What You Are Actually Running/1. D04_vLLM_Hardening_Checklist_and_Flag_Reference.docx

24 kB

1 - What You Are Actually Running/1. D05_Authenticating_Gateway_Guide_TLS_Per_Client_Keys_Endpoint_Allowlist_Limits.d.docx

22 kB

1 - What You Are Actually Running/1. D06_Network_Isolation_Register_Firewall_Policy_Egress_Listeners_and_Remote_Acces.docx

22 kB

1 - What You Are Actually Running/1. D07_Model_Intake_and_Manifest_Policy_Formats_Pinning_Signing_Licences.docx

33 kB

1 - What You Are Actually Running/1. D08_Secrets_Prompt_Data_and_Retention_Policy_with_a_Model_Disposal_Procedure.doc.docx

31 kB

1 - What You Are Actually Running/1. D09_Monitoring_Alerting_and_Detection_Runbook.docx

38 kB

1 - What You Are Actually Running/1. D10_Advisory_Watch_and_Controlled_Upgrade_Procedure.docx

35 kB

1 - What You Are Actually Running/1. D11_Incident_Playbooks_for_a_Model_Host_and_Incident_Note_Template.docx

27 kB

1 - What You Are Actually Running/1. D12_Framework_Crosswalk_and_Auditor_Evidence_Pack_Index.docx

36 kB

1 - What You Are Actually Running/2. Section 1 Intro A Model Server Is a Server.mp4

14 MB

1 - What You Are Actually Running/3. Why Organisations Self-Host Open-Weight Models — and What They Take On.mp4

24 MB

1 - What You Are Actually Running/4. The Assets Weights, Prompts, Outputs, Logs, Compute and the API Itself.mp4

23 MB

1 - What You Are Actually Running/5. Trust Boundaries of an Inference Host, Drawn Once.mp4

25 MB

1 - What You Are Actually Running/6. What Researchers Found on the Internet Exposed Ollama Servers, 2025-2026.mp4

22 MB

1 - What You Are Actually Running/7. The OWASP Top 10 for LLM Applications 2026, Read as an Operator.mp4

27 MB

1 - What You Are Actually Running/8. NIST AI RMF, AI 600-1 and the Joint Deploying AI Systems Securely Guidance.mp4

28 MB

1 - What You Are Actually Running/9. [LAB] Tour of the Bench CPU, Memory, No GPU — and Why That Is Enough.mp4

10 MB

1 - What You Are Actually Running/10. Section 1 Wrap Your Threat Model on One Page.mp4

7.0 MB

2 - The Default Install, Measured/11. Section 2 Intro Before You Fix It, Look at It.mp4

13 MB

2 - The Default Install, Measured/12. [LAB] Installing Ollama the Documented Way and Listing What It Opens.mp4

25 MB

2 - The Default Install, Measured/13. [LAB] The Ollama API Surface Every Endpoint, Which Ones Change State.mp4

22 MB

2 - The Default Install, Measured/14. [LAB] Ollama in Docker as Shipped Root User, 0.0.0.0 Inside, What -p Publishes.mp4

22 MB

2 - The Default Install, Measured/15. [LAB] vLLM CPU Container as Shipped Start-up, Warm-up and the Endpoint Map.mp4

25 MB

2 - The Default Install, Measured/16. [LAB] Choosing Lab Models by Licence Apache-2.0, MIT and the Ones That Are Not.mp4

22 MB

2 - The Default Install, Measured/17. [LAB] Opening llm-host The Repo and a verify.sh Where Every Check Fails.mp4

27 MB

2 - The Default Install, Measured/18. Section 2 Wrap The Baseline Scoreboard.mp4

7.1 MB

3 - Hardening Ollama on the Host/19. Section 3 Intro One Binary, One Port, No Password.mp4

7.3 MB

3 - Hardening Ollama on the Host/20. Does Not Require Authentication What the Docs Say and What It Means for You.mp4

12 MB

3 - Hardening Ollama on the Host/21. [LAB] Bind Address OLLAMA_HOST, Loopback and Proving It With ss.mp4

20 MB

3 - Hardening Ollama on the Host/22. [LAB] CORS and OLLAMA_ORIGINS What the Browser Can Reach.mp4

20 MB

3 - Hardening Ollama on the Host/23. DNS Rebinding and Browser-Reachable Local APIs — Why Host Checks Exist.mp4

19 MB

3 - Hardening Ollama on the Host/24. [LAB] The ollama Service User, Model Directory Ownership and File Modes.mp4

23 MB

3 - Hardening Ollama on the Host/25. [LAB] systemd Sandboxing for ollama.service ProtectSystem, NoNewPrivileges.mp4

29 MB

3 - Hardening Ollama on the Host/26. [LAB] Scoring the Unit With systemd-analyze security, Before and After.mp4

20 MB

3 - Hardening Ollama on the Host/27. [LAB] Resource Limits That Matter Queue, Parallelism, Loaded Models, Context.mp4

24 MB

3 - Hardening Ollama on the Host/28. [LAB] Turning Off What You Do Not Use Cloud Features, Debug Request Logging.mp4

20 MB

3 - Hardening Ollama on the Host/29. Section 3 Wrap Ollama on the Host, Locked to Loopback.mp4

11 MB

4 - Hardening Ollama in Containers/30. Section 4 Intro The Container Is Not the Boundary Until You Make It One.mp4

15 MB

4 - Hardening Ollama in Containers/31. [LAB] Running Ollama as a Non-Root User With a Correctly Owned Volume.mp4

24 MB

4 - Hardening Ollama in Containers/32. [LAB] Read-Only Root Filesystem, tmpfs and Dropping Every Capability.mp4

21 MB

4 - Hardening Ollama in Containers/33. [LAB] no-new-privileges, pids, Memory and CPU Limits for a Runaway Prompt.mp4

19 MB

4 - Hardening Ollama in Containers/34. [LAB] Publishing on 127.0.0.1 Only and Checking It From Another Host.mp4

18 MB

4 - Hardening Ollama in Containers/35. [LAB] An Internal Docker Network With No Route Out.mp4

29 MB

4 - Hardening Ollama in Containers/36. [LAB] Pinning the Image by Digest Instead of latest.mp4

26 MB

4 - Hardening Ollama in Containers/37. [LAB] Scanning the Ollama Image SBOM With Syft, Findings With Grype.mp4

23 MB

4 - Hardening Ollama in Containers/38. Rootless Docker and User-Namespace Remapping When to Use Which.mp4

24 MB

4 - Hardening Ollama in Containers/39. Section 4 Wrap A Container You Would Put on a Shared Host.mp4

11 MB

5 - Hardening vLLM/40. Section 5 Intro More Knobs, More Endpoints.mp4

13 MB

5 - Hardening vLLM/41. The vLLM Security Page, Read Line by Line.mp4

26 MB

5 - Hardening vLLM/42. [LAB] --host, --port and Why You Never Rely on the Default Bind.mp4

19 MB

5 - Hardening vLLM/43. [LAB] --api-key and Its Real Scope Measuring 401 Versus 200 Per Endpoint.mp4

23 MB

5 - Hardening vLLM/44. [LAB] Closing the Extras --disable-fastapi-docs and CORS With --allowed-origins.mp4

20 MB

5 - Hardening vLLM/45. [LAB] trust_remote_code Stays Off What the Flag Loads and How to Check It.mp4

18 MB

5 - Hardening vLLM/46. [LAB] Media and SSRF Controls Local Media Paths, Allowed Domains, Size Limits.mp4

28 MB

5 - Hardening vLLM/47. [LAB] Per-Request Caps max-model-len, max-num-seqs, VLLM_MAX_N_SEQUENCES.mp4

25 MB

5 - Hardening vLLM/48. [LAB] Running the vLLM Container as a Non-Root User, Read-Only Where It Can Be.mp4

23 MB

5 - Hardening vLLM/49. Dangerous Switches Dev Mode, Runtime LoRA, Insecure Serialization, Plugins.mp4

14 MB

5 - Hardening vLLM/50. Multi-Node vLLM Why Inter-Node Traffic Must Live on an Isolated Network.mp4

24 MB

5 - Hardening vLLM/51. Section 5 Wrap vLLM Behind Its Own Flags.mp4

13 MB

6 - The Authenticating Gateway/52. Section 6 Intro Authentication Lives in Front of the Model.mp4

7.1 MB

6 - The Authenticating Gateway/53. Gateway Patterns Reverse Proxy, API Gateway, LLM Proxy — What Each Adds.mp4

22 MB

6 - The Authenticating Gateway/54. [LAB] nginx in Front of Ollama TLS With a Private CA.mp4

27 MB

6 - The Authenticating Gateway/55. [LAB] API Keys at the Gateway One Key Per Client, Checked Before the Model.mp4

25 MB

6 - The Authenticating Gateway/56. [LAB] The Endpoint Allowlist Inference In, Model Management and metrics Out.mp4

28 MB

6 - The Authenticating Gateway/57. [LAB] Rate Limiting and Connection Limits Watching 429 Arrive.mp4

22 MB

6 - The Authenticating Gateway/58. [LAB] Body Size, Timeouts and Streaming Responses Through the Proxy.mp4

22 MB

6 - The Authenticating Gateway/59. [LAB] The Same Gateway in Front of vLLM, With Its Unprotected Paths Closed.mp4

26 MB

6 - The Authenticating Gateway/60. [LAB] An LLM Proxy Layer Per-Client Virtual Keys and Budgets With LiteLLM.mp4

32 MB

6 - The Authenticating Gateway/61. Section 6 Wrap Every Request Now Has a Name.mp4

13 MB

7 - Network Isolation and Egress/62. Section 7 Intro What the Model Server May Talk To.mp4

13 MB

7 - Network Isolation and Egress/63. [LAB] A Host Firewall With nftables Default Deny, One Port In.mp4

27 MB

7 - Network Isolation and Egress/64. [LAB] Egress Control The Model Server Has No Business on the Internet.mp4

23 MB

7 - Network Isolation and Egress/65. [LAB] Pull Through a Staging Step Fetch, Check, Promote Into the Isolated Host.mp4

24 MB

7 - Network Isolation and Egress/66. [LAB] Finding Every Listening Socket, Including the Ones You Did Not Ask For.mp4

26 MB

7 - Network Isolation and Egress/67. Remote Access Without Exposure VPN, Bastion and Private Overlay Networks.mp4

12 MB

7 - Network Isolation and Egress/68. Section 7 Wrap The Host From the Outside.mp4

10 MB

8 - Model Supply Chain and Integrity/69. Section 8 Intro A Model File Is Code-Adjacent.mp4

13 MB

8 - Model Supply Chain and Integrity/70. Pickle, safetensors, GGUF What Each Format Can and Cannot Do When Loaded.mp4

30 MB

8 - Model Supply Chain and Integrity/71. [LAB] Reading a safetensors Header by Hand.mp4

18 MB

8 - Model Supply Chain and Integrity/72. [LAB] Scanning Model Files modelscan, picklescan, fickling and a Test Fixture.mp4

28 MB

8 - Model Supply Chain and Integrity/73. torch.load, weights_only and the 2.6 Default Change.mp4

20 MB

8 - Model Supply Chain and Integrity/74. [LAB] Pinning Hugging Face Downloads to a Full Commit SHA With hf download.mp4

26 MB

8 - Model Supply Chain and Integrity/75. [LAB] Offline Mode at Serve Time HF_HUB_OFFLINE and a Local Model Path.mp4

21 MB

8 - Model Supply Chain and Integrity/76. [LAB] Verifying Ollama Blobs Against Their sha256 Digests.mp4

24 MB

8 - Model Supply Chain and Integrity/77. [LAB] Signing a Model With OpenSSF model_signing and Verifying It Before Serve.mp4

26 MB

8 - Model Supply Chain and Integrity/78. [LAB] The Model Manifest Source, Revision, Digest, Signature and Licence.mp4

26 MB

8 - Model Supply Chain and Integrity/79. Section 8 Wrap Nothing Loads That Is Not in the Manifest.mp4

12 MB

9 - Secrets, Prompts and Data/80. Section 9 Intro The Data That Flows Through the Model.mp4

12 MB

9 - Secrets, Prompts and Data/81. [LAB] Gateway Keys and HF Tokens Out of Compose Docker Secrets, systemd Creds.mp4

28 MB

9 - Secrets, Prompts and Data/82. [LAB] What the Servers Log About Prompts, and Keeping It That Way.mp4

24 MB

9 - Secrets, Prompts and Data/83. Hidden Context Exposure Why Secrets Never Go in a System Prompt.mp4

12 MB

9 - Secrets, Prompts and Data/84. [LAB] Cache and Model Directories Permissions, Ownership and What Lives There.mp4

22 MB

9 - Secrets, Prompts and Data/85. Retention, Disposal and Deleting a Model Properly.mp4

12 MB

9 - Secrets, Prompts and Data/86. Section 9 Wrap Where Every Secret on the Host Lives.mp4

9.6 MB

10 - Logging, Monitoring and Detection/87. Section 10 Intro You Cannot Defend What You Cannot See.mp4

14 MB

10 - Logging, Monitoring and Detection/88. [LAB] Structured JSON Access Logs at the Gateway.mp4

22 MB

10 - Logging, Monitoring and Detection/89. [LAB] Scraping vLLM metrics Privately With Prometheus.mp4

23 MB

10 - Logging, Monitoring and Detection/90. [LAB] A Grafana Board for Requests, Errors, Queue Depth and Tokens.mp4

27 MB

10 - Logging, Monitoring and Detection/91. [LAB] Alert Rules That Mean Something 401 Bursts, 429 Storms, Unknown Models.mp4

19 MB

10 - Logging, Monitoring and Detection/92. [LAB] Falco on the Model Container A Shell or a Write Where None Should Be.mp4

26 MB

10 - Logging, Monitoring and Detection/93. [LAB] Detecting Model Changes A Scheduled Manifest Check That Pages You.mp4

18 MB

10 - Logging, Monitoring and Detection/94. Section 10 Wrap The Host Now Talks Back.mp4

8.7 MB

11 - Patching and Vulnerability Management/95. Section 11 Intro These Projects Ship Fast — So Do Their Advisories.mp4

7.2 MB

11 - Patching and Vulnerability Management/96. Reading the Advisory History Ollama, vLLM and llama.cpp by Class of Flaw.mp4

26 MB

11 - Patching and Vulnerability Management/97. [LAB] Querying OSV and GitHub Advisories for the Versions You Run.mp4

26 MB

11 - Patching and Vulnerability Management/98. [LAB] A Controlled Upgrade New Digest, Smoke Test, verify.sh, Rollback.mp4

22 MB

11 - Patching and Vulnerability Management/99. GPU Hosts NVIDIA Container Toolkit Bulletins and Driver Hygiene (Conceptual).mp4

25 MB

11 - Patching and Vulnerability Management/100. Section 11 Wrap A Patch Rhythm You Can Keep.mp4

9.0 MB

12 - Operating, Responding and Proving It/101. Section 12 Intro From Hardened to Run Well.mp4

12 MB

12 - Operating, Responding and Proving It/102. [LAB] Incident Drill Your Model API Was Reachable From the Wrong Network.mp4

25 MB

12 - Operating, Responding and Proving It/103. [LAB] Incident Drill An Unexpected Model Appeared on the Host.mp4

21 MB

12 - Operating, Responding and Proving It/104. Mapping llm-host to NIST AI RMF, ISO IEC 42001, UK AI Cyber Security Code.mp4

27 MB

12 - Operating, Responding and Proving It/105. EU AI Act for People Who Self-Host Deployer, Provider and the Dated Milestones.mp4

26 MB

12 - Operating, Responding and Proving It/106. MITRE ATLAS Mitigations You Have Already Implemented.mp4

25 MB

12 - Operating, Responding and Proving It/107. [LAB] Rebuild From Scratch The Whole Hardened Stack From One Repository.mp4

29 MB

12 - Operating, Responding and Proving It/108. [LAB] Final verify.sh Run and the Evidence Pack for Your Auditor.mp4

24 MB

12 - Operating, Responding and Proving It/109. Section 12 Wrap What You Now Run, and What Still Needs a Human.mp4

12 MB