Training · Cyber Defence
Assisted Threat Hunting
Put local AI to work on real attack data. Turn hunt ideas into tested queries, find anomalies at scale and build hunting agents whose findings you can verify.
- Advanced
- 40 hours
- 5 days
- 5-day bootcamp
- AI-assisted
About the course
Hunters now face more telemetry than any team can read. Large language models can turn a hunt idea into a working query and summarise thousands of events in seconds. Used carelessly, they invent field names, miss evidence and report attacks that never happened.
This bootcamp teaches hunters to use AI without trusting it blindly. Every lab runs on locally hosted open-weight models, so logs, prompts and findings never leave the lab, and every AI finding is checked against the raw data.
Who should attend
- Threat Hunting graduates
- SOC L2 and L3 analysts and threat hunters
- Detection engineers and SIEM content owners
- Threat intelligence analysts
- Security automation and SOAR engineers
What the course covers
Six topics in focus
01
LLMs for hunters
Hypotheses from intel and ATT&CK, and where models fail.
02
Natural language to query
Plain-English hunt questions turned into SPL, KQL and Sigma.
03
Machine learning at scale
Anomaly detection, clustering and AI summaries of large log sets.
04
Agentic hunting
Agents that pivot across data, with human checkpoints.
05
Adversarial AI
Prompt injection hidden in log fields to mislead AI hunters.
06
Hunt to detection
AI-written Sigma rules, tested, tuned and measured.
Tools and technologies
- Splunk
- Model Context Protocol
- Ollama
- vLLM
- Splunk AI Toolkit
- Sigma
- MISP
Course outline
5 days, one capstone
- Day 1 LLMs for threat hunters Where AI fits in a hunt, how it fails, and how to check its output.
- Day 2 Natural language to query Connecting a model to Splunk and turning hunt questions into tested queries.
- Day 3 Machine learning and AI at scale Anomalies, clustering and AI-built timelines, each proven against raw events.
- Day 4 Agentic hunting Hunting agents with guardrails, and defending them against prompt injection.
- Day 5 Hunt to detection AI-written detections, measuring AI results, and the capstone.
Capstone
AI-assisted versus manual
Teams hunt with their own co-pilot and agent, turn one finding into a tested Sigma rule, and compare speed and accuracy with a manual hunt.
Outcomes
You will be able to
- Run local LLMs safely on security telemetry
- Turn hunt hypotheses into tested SPL with an AI co-pilot
- Use machine learning to find anomalies and cluster events at scale
- Build a hunting agent with guardrails
- Generate, test and measure AI-written detections
Before you join
Prerequisites
- Threat Hunting or equivalent hands-on hunting experience
- Working knowledge of Splunk SPL and MITRE ATT&CK
- Comfortable writing Python scripts
- No prior AI or machine-learning experience needed
Get the full syllabus
Batches start every month from January 2027. Send us an enquiry and we'll email you the detailed brochure and the next batch dates.