
AI Native Engineer Track
Ship two production-grade AI projects in eight weeks.
The second rung. Model selection and orchestration, prompt engineering as a professional discipline, retrieval-augmented generation, agents and tool use, the Audit Discipline with the open-source Nexuscone primitive, and testing AI-built code with the same rigour as any other production system. Capstone delivers two deployed AI applications with full architectural decision records.
- Price
- £1,499
- Length
- 8 weeks
- Prerequisite
- Foundation or equivalent
- Cohort size
- Capped at 20
- Time commitment
- 15-20 hours per week
What you could build
AI Native graduates ship two production-grade AI projects. This is the kind of system one of them could be.
Live
Production-grade AI project
Curriculum
- Week 1 · Model selection and orchestration
- The current range of models. Cost per token, cost per task, cost per user. LLM routers including the Aperintelligence AI Gateway pattern. Multi-provider fallback chains using Marktrader's Lumen cascade as the worked example. Caching to reduce cost.
- Week 2 · Prompt engineering as a discipline
- The Aperintelligence Prompting Protocol at depth. The Personal Genome and the Founder Genome. Anti-cliché output discipline. Structured output: JSON schemas, function calling, tool use. The honour code on AI-assisted output.
- Week 3 · Retrieval-augmented generation
- What an embedding is. Vector databases (pgvector, Pinecone, Weaviate, Qdrant). Chunking strategies. Hybrid search as the default. Citation patterns. Worked example: the Aperintelligence AI Gateway's retrieval pattern.
- Week 4 · Agents and tool use
- The difference between an LLM call and an agent. MCP servers, function calling, multi-agent orchestration. The runtime cost and runtime risk of agentic systems. When NOT to use an agent (most of the time).
- Week 5 · Audit Discipline with Nexuscone
- Why audit trails are mandatory for production AI. The cryptographic primitives in plain English. The Nexuscone library walkthrough. Wiring Nexuscone into any application. Cryptographic timestamp anchoring with OpenTimestamps.
- Week 6 · Testing AI-built code
- Why AI-generated code needs more testing, not less. Unit testing, integration testing, evaluation suites for AI behaviour. Regression testing when you swap models. Test-driven development with AI in the loop.
- Week 7 · Capstone build
- Two production-grade AI projects, designed by the student, deployed publicly, with full architectural decision records. Mid-cohort review and one-to-one mentor sessions on the capstone.
- Week 8 · Career engine for AI-native engineers
- Presenting AI-assisted shipping evidence in interviews. What hiring pipelines screen for now. Day-rate contracting versus permanent employment. Building a public portfolio that compounds.