What Is a Large Language Model?
Short Answer
A large language model (LLM) is a neural network trained on a huge amount of text to predict the next token, a piece of a word. Repeating that prediction lets it write answers, summaries and code. GPT, Claude, Gemini, Llama and Mistral models are all large language models.
- Taught in lesson 2
- In 5 real job ads
Why It Matters
Large language models are behind the AI tools that changed how people write, search and code. Most new AI products and many new AI jobs are built around them, so understanding how they work, what they cost and where they fail is now a core skill for AI engineers and useful for everyone around them.
How It Works
An LLM is first pretrained on text from the web, books and code, which teaches it language and a broad base of knowledge. It is then fine-tuned to follow instructions and tuned with human feedback to be helpful and safe. Almost all of them use the transformer architecture.
An LLM has no built-in access to your company data or to events after its training cutoff, and it can state wrong things confidently. Most AI engineering work is about wrapping the model with retrieval, tools, checks and evaluation so it becomes reliable.
An LLM has no database of facts it looks up. Its knowledge is spread across its weights as patterns learned from text, and it has a training cutoff after which it knows nothing unless you give it information in the prompt. This is why applications add retrieval for company or recent data, tools for actions such as calculations, and evaluation to check answers. Models differ in size, speed, cost and strengths, so teams choose per task.
Large Language Model in Practice
Everyday Example
When you ask an assistant to explain a contract clause in plain words, a large language model generates the explanation one token at a time.
Industry Example: Pharma
An LLM can draft a plain-language summary of a clinical trial result for patients, which a medical writer then checks against the source data.
Where You Meet It
- Chat assistants, writing tools and coding assistants.
- API calls in AI engineer work.
- Job ads asking for LLM, GenAI or foundation model experience.
- Product features that summarise, draft, classify or answer questions.
Key Ideas Around It
- Token
- The piece of text an LLM reads and writes.
- Context window
- How much text it can consider at once.
- Training cutoff
- The date after which the model learned nothing new.
- Inference
- Using the trained model to generate an answer.
A Common Mix-Up
A bigger model is not always better for a task. Smaller models can be faster, cheaper and just as good on a narrow job; only a test on your own task tells.
Check YourselfWhy might an LLM give an outdated answer about a recent event?
It only knows what was in its training data up to its cutoff, unless the application gives it recent information in the prompt.
Seen in Real Job Ads
Lines quoted word for word from live job postings.
Pfizer
Senior Manager, LLMOps & AI Integration Engineer
Hands‑on with GenAI/LLM ecosystems (e.g., tool/function calling, structured outputs, prompt orchestration) and RAG components (embeddings, retrievers, rerankers, vector stores).
Technergetics
AI/ML Engineer III, Rome, NY
Ability to design and run LLM evaluation, including task-specific benchmarks, golden datasets, LLM-as-judge methods, and tracing and observability tooling (LangSmith, Langfuse, Arize, Weights & Biases)
Angi
Principal Product Manager, AI Platform, Denver, CO
Hands-on familiarity with LLM gateways/proxies (e.g., LiteLLM), evaluation and observability tooling (e.g., Langfuse), and model serving frameworks (e.g., KServe, SageMaker, Bedrock, Ray).
Learn It on ai2hired
S0 · #2What AI Really Is: AI vs Machine Learning vs LLMs
This free lesson teaches Large Language Model from zero, with the job ads that ask for it.
Questions People Ask
Does an LLM understand what it says?
It builds rich internal representations of language, but it produces text by predicting likely tokens. That is why it can sound sure while being wrong.
What makes a language model large?
The number of parameters, often billions, and the amount of text it was trained on, often trillions of tokens.
Related Terms
Learn AI From Zero to Hired
Free lessons on YouTube, each one built from what real AI job ads ask for.