tecnovy AI Suite

Building LLM Applications

At the bottom sits a call to a model API that catches errors and records tokens and response time. On top of it you place a versioned prompt that sets the behaviour. A JSON schema gives the answer a shape your code can process. A tool call lets the model request one of your functions, which your code runs only after checking it. Tests come last and show whether each layer does its job. You build these layers in Python, one after the other.

Attendance certificate4.8/5 on ProvenExpert

What this module delivers.

Developers, data scientists and tech leads use tecnovy's LLM Engineering to build language models into software without tying it to one vendor. Afterwards your team picks models by quality, cost and latency and checks LLM features with its own tests.

How the training runs

Method: Short theory blocks, then Python exercises on one continuous project
Basis: Vendor-neutral tecnovy curriculum, as of November 2026
Outcome: tecnovy certificate of participation, no examination

Dates & booking

Choose a date that fits

No public dates are scheduled yet.

New dates are published regularly. Join the waitlist or plan an in-house course for your team.

Request in-house training

Fit

Who this module is designed for

Typical roles

You develop software and are asked to add LLM features to existing applications.
You work as a data scientist or ML engineer and want to build language models into your own applications.
You are a tech lead and assess LLM initiatives by model choice, cost and risk.

Prerequisites

No prerequisites

You can start right away. What you need: programming experience in Python, basic knowledge of REST APIs and JSON.

Consider instead AI Agents Foundations Without Python experience: you assess processes by benefit, effort and risk and build first AI agents with no-code tools.

Curriculum

LLM Engineering in Detail

Five parts in the order an LLM application comes together: from the model itself through the API call, prompts and structured outputs to function calls and tests. Retrieval from company documents and multi-step agents belong to the tecnovy courses RAG and Developing AI Agents.

01How LLMs work

Before writing the first line of code, you clarify what happens inside the model and what it costs.

Tokens, context windows and sampling parameters such as temperature
Model families compared: strengths, cost and latency
Criteria for choosing a model for a specific task

02Working with model APIs

The API call is the bottom layer everything else builds on.

Requests with system, user and assistant roles
Streaming: receiving the answer piece by piece
Errors, timeouts and retries
Measuring and reducing cost and latency

03Prompt design for applications

A prompt inside an application is code and gets treated as such.

System prompts and examples in the prompt
Templates and versioning
Structured outputs with JSON schemas and checking them in code

04Tool use and function calls

Here the model requests functions of your application while your code stays in control.

Describing functions so that the model requests them correctly
The round trip of one call: request, execution, result returned
Checking arguments before a function runs

05Checking quality and risks

Finally you check whether your application delivers what it should.

Test cases and evaluations for LLM outputs
Hallucinations: plausible-sounding but invented content
Prompt injection: instructions smuggled in through inputs
Data protection: which data goes to a model provider

Outcome

What you will be able to do afterwards

01
You choose a model by quality, cost, latency and context window and justify the choice.
02
You integrate a model API into your code with roles, streaming, error handling and retries.
03
You measure the token use and response time of your calls and derive measures to reduce both.
04
You write system prompts with examples as templates and version them like code.
05
You define JSON schemas for structured outputs and check every answer against them before your code uses it.
06
You let the model request functions of your application and check the arguments before your code runs the call.
07
You write test cases and evaluations that check LLM outputs against expected results.
08
You test your application for invented content (hallucinations), injected instructions (prompt injection) and its handling of personal data.

tecnovy Teilnahmezertifikat

Confirms your attendance and the curriculum covered in the training. Issued by tecnovy, shareable on LinkedIn.

Open the Certificate Showroom tecnovy →

Offer scope

What is included

Included

Project repository with sample codeIt holds the code of the course's continuous project, which you can start from for your own LLM applications after the training.

Why tecnovy

What you get on top with us

No Slideshow, Hands-On!

Promised: no PowerPoint marathon. We work in groups, tie theory to practice, and you get real project examples from our experienced trainers plus the exchange with like-minded people.

Certificate Showroom

Get your certificate of participation and, if you have one, add your exam certificate from E-Learning. Fully automated, beautifully designed. Just for you, only at tecnovy.

Flexible Date Change

If you are not able to attend the course, you can rebook your training free of charge up to one week before the start of the training.

Attend Twice, Pay Once

You are welcome to attend the training online again within a year as a refresher.

Learn from Experts

We always guarantee you the use of didactically and methodically first-class qualified trainers who draw their knowledge from training experience as well as professional practical and project experience.

FAQs

Frequently asked questions

01Which models does the course work with?

tecnovy's LLM Engineering works vendor-neutrally with common commercial and open models through their APIs. You build prompt templates, JSON schemas, function calls and tests so that you can carry them over to another provider.

02Do I need prior knowledge of machine learning?

No. For tecnovy's LLM Engineering, programming experience in Python and basic knowledge of REST APIs and JSON are enough, because the course has you call models through their APIs rather than train them.

03Is the course suitable for Java or .NET developers?

Yes, provided you also have programming experience in Python, because the exercises in tecnovy's LLM Engineering run in Python. You can then carry the patterns themselves, from prompt template to test case, over to other languages.

04Do structured outputs guarantee a valid answer?

Not in every case. Anthropic and OpenAI state that answers produced with structured outputs follow a supplied JSON schema, but both name exceptions such as a refusal by the model or a reached token limit (as of November 2026). In tecnovy's LLM Engineering your code therefore checks every answer before using it.

05How far does the course go with tool use?

tecnovy's LLM Engineering covers the single tool call, which some providers call function calling: the model requests a function of your application, your code checks the arguments, runs it and returns the result. Agents that chain such calls into multi-step flows are what you build in the tecnovy workshop Developing AI Agents.

06Do the exercises use real model APIs?

Yes, every exercise in tecnovy's LLM Engineering addresses a language model through its API and needs API access with a key for that. Calls use tokens; Anthropic bills input and output tokens separately (as of November 2026). You measure usage in the course.

07Does the course make my application safe against prompt injection?

No, tecnovy's LLM Engineering does not promise that. You learn to test your application with your own test cases for injected instructions and invented content, and to decide which data you send to a model provider at all.

08What proof of completion does LLM Engineering give?

The tecnovy certificate of participation, which you receive with at least 80% attendance. The tecnovy LLM Engineering training includes no examination and no certification from a model vendor.

09Which course builds on LLM Engineering?

The next step at tecnovy is the RAG training: there you connect an LLM application to company documents, through search and answers with source references. Retrieval, finding the relevant content, is deliberately not part of LLM Engineering.

What does your training at tecnovy look like?

LLM EngineeringBuilding LLM Applications

No dates yet

We will let you know as soon as the next date is scheduled.