Skip to content
AI Agency

The AI Forge · Foundation · F1

AI Engineering Foundations

Context engineering, model APIs, tool use, evaluation, cost and security basics

Duration
2 days · 9:00–17:00 each day (13h contact, incl. breaks + lunch)
Participant level
Intermediate: regular AI users
Format
Instructor-led labs, build-along
Participants
8 included · 12 maximum (flexible group size)
Prerequisite
No prior module, programming proficiency required
Ecosystem
Agnostic: any model provider or stack
Price
EUR 7,000 (EUR 250 per extra participant)
Discovery Session
On request

Amounts in EUR, excl. VAT.

A trainer presenting to participants working on laptops in a bright training roomai-generated

What participants will be able to do

  • Set up and run LLM inference pipelines end to end
  • Evaluate and compare foundation models for a given use case
  • Build a prompt engineering library with versioning and eval
  • Implement basic safety patterns and output guardrails

Tools & resources

PythonOpenAI API / Anthropic APILangChainHugging FaceJupyter

What changes after this module

Engineers can call any model API reliably, apply context engineering over prompt-craft, evaluate outputs systematically, and ship a cost- and safety-aware LLM endpoint, the durable base every other Forge module builds on.

Who should attend

Engineers and data professionals. Intermediate: regular AI users.

Programme

This agenda is indicative. Content, sequencing, and examples are adapted to your team's context, tools, and objectives.

Before you start

  • A laptop with a working development environment and admin rights
  • Proficiency in at least one general-purpose language (Python and/or JavaScript/TypeScript)
  • A code editor/IDE with an approved AI coding assistant enabled
  • An API key for an approved model provider (issued by the company)
  • A Git repository and command-line comfort
  • No prior LLM experience required; bring one real task you want an endpoint to perform
Day 19:00–10:30How models actually work: Tokenisation, context windows, sampling and why models behave probabilistically: the mental model that outlasts any single model release Lab: Probe a model’s limits: run controlled prompts varying temperature and context, and chart where output quality breaks.
10:30–10:45 · ☕ Coffee break
Day 110:45–12:30Model APIs the right way: Authentication, streaming, retries, timeouts, structured outputs and function/tool calling: the integration contract that stays stable across providers Lab: Build a resilient API client with streaming, retries and structured-output parsing against a live model endpoint.
12:30–13:30 · 🍽 Lunch break
Day 113:30–15:30Context engineering: The 2026 successor to prompt engineering: designing everything the model sees each call: system, inputs, retrieved context, tool defs, memory: and budgeting the token window Lab: Take a failing task and fix it purely by re-engineering context and token budget; measure the before/after quality.
15:30–15:45 · ☕ Coffee break
Day 115:45–17:00Lab: build an LLM endpoint: Assemble the pieces into a working service with clean prompt/context management Lab: Ship a working LLM-powered API endpoint with versioned prompts and structured output.
Day 29:00–10:30Evaluation-driven development: Why you build the eval harness before you optimise; offline vs online eval; LLM-as-judge; regression detectionEngineers without evals ship regressions silently Lab: Write an eval harness (assertion + LLM-as-judge) for your endpoint and wire a regression check.
10:30–10:45 · ☕ Coffee break
Day 210:45–12:30Cost & latency engineering: Token economics, caching, model selection by task, prompt compression: durable levers independent of price sheets Lab: Instrument cost and latency, add caching and a task-appropriate model choice, and quantify the saving.
12:30–13:30 · 🍽 Lunch break
Day 213:30–15:30Security basics: Prompt injection, sensitive-data disclosure, untrusted output handling, and the defence-in-depth mindset (least privilege, filtering, human approval) Lab: Attack your own endpoint with injection payloads, then add input/output filtering and re-test.
15:30–15:45 · ☕ Coffee break
Day 215:45–17:00Deploy & wrap: Deploy, add basic logging, and review against the eval harness Lab: Deploy the endpoint, confirm logging and evals run, and demo it to the room.

Deliverables

  • Working LLM-powered API endpoint (versioned prompts, structured output)
  • Reusable eval harness with a regression check
  • Cost/latency instrumentation + caching
  • Input/output safety filter

Interested in running this module for your team? Get in touch and we'll tailor the format, dates, and delivery to your context.

Request this training

Your trainer

Practising AI engineer; ships LLM applications to production; strong evaluation discipline

Discovery Session

We also offer a Discovery Session: a 3-hour introduction to the core concepts, with a hands-on exercise, on request, for groups of 5 to 15.

3 hours · From €3,500 · Groups of 5 to 15 - On request

Request a Discovery Session

Part of a bigger path

  • EnterpriseEUR 115,000 · up to 30 training days
See the programmes

Frequent questions

Can modules be taken individually?

Yes. Every module stands alone at a fixed price, and every module counts toward a programme if you continue.

Where does training happen?

At your premises or remote, on your dates, for private cohorts. Open sessions run on a fixed monthly calendar.

Which AI tools do you train on?

Yours. Every module ships in four ecosystem editions and runs its exercises on your real stack.

Who delivers?

Inforca's senior consultants and trainers. Flagged modules and the executive track are delivered at senior-expert level.

Discuss this module in a First Call: fit, dates, and the path around it.

Book a First Call

No commitment · our team responds within one business day