Career Path
AI Research (The Foundation)
Builds the deep understanding of models that research roles start from.
- 103 Free Lessons
- 10 Seasons
- 10 Interview Questions
What The Job Is
AI research roles start from a deep understanding of how models work: the mathematics behind them, how neural networks and transformers learn, and how to measure whether a new idea really helps.
This path builds that foundation, from statistics and classical machine learning to deep learning, training and customizing models, and reading research papers with a critical eye.
At a Glance
- 6 of 103lessons out now
- 10interview questions
- 8key skills explained
Job Titles to Search
- Research Scientist
- Research Engineer
Your First Lessons
The Job in Detail
Why Companies Hire for It
Research teams push what models can do: new training methods, new architectures and better ways to measure and understand behaviour. Large AI labs, big technology companies and some biotech and pharma companies hire researchers and research engineers to create the methods that products will use later.
A Typical Week
- 1Read recent papers and reproduce results to see what really works.
- 2Design experiments that test one idea about how models learn or behave.
- 3Write training and evaluation code, often on large clusters of GPUs.
- 4Analyse results with care, including when an idea does not work.
- 5Write up findings for colleagues or publication and discuss them in reading groups.
Tools and Terms You Will Meet
- PyTorch
- The framework most research code is written in.
- Linear algebra
- The math of vectors and matrices that every model is built on.
- Gradient descent
- The method models use to learn from their errors.
- Transformers
- The model design behind today's language models.
- arXiv
- The site where most AI research papers appear first.
Who Moves Into This Job
Research scientists usually hold a PhD in machine learning, computer science, physics, math or a related field. Research engineers, who build the systems research runs on, more often come from strong software engineering with deep interest in models. This path on ai2hired builds the foundation both start from.
Industry Example: Pharma
In pharma and biotech, AI research focuses on models for biology and chemistry: protein structure, molecule generation, predicting how a compound behaves, and reading medical images. Teams combine machine learning researchers with biologists and chemists.
A Common Mix-Up
AI research is not the same as building AI products. Researchers create new methods and test them; engineers turn proven methods into products. Most AI jobs are engineering jobs, and many research teams hire research engineers alongside scientists.
Is This Job for You?
A good fit if
- You enjoy math and want to understand why models work.
- You are patient with experiments that fail.
- You like reading papers and writing precisely.
Maybe not if
- You want to ship products to users quickly.
- You prefer clear tasks to open questions.
- You do not want long study, often to doctoral level.
Where the Job Leads
Researchers grow into senior and principal scientist roles, research lead, or move to applied science and engineering roles where their depth is rare. Research engineers often become ML or infrastructure leads.
How to Get In
- 1Build strong foundations in Python, linear algebra, probability and calculus.
- 2Implement a small neural network and a small transformer from scratch, and explain each part.
- 3Reproduce one published result and write up what matched and what did not.
What Employers Ask For
Real lines from real job ads for this role, each with the lesson that teaches it.
GSK
AIML Software Engineer, AI for Science, London
Basic understanding of at least one major deep learning framework (PyTorch, JAX, TensorFlow).
Taught InS7 · #74PyTorch Fundamentals
Merck (MSD)
2027 Nonclinical Drug Safety Data Scientist Intern, Boston, MA or West Point, PA
Familiarity with statistical learning methods, such as supervised and unsupervised modeling
Taught InS6 · #68Clustering and Dimensionality Reduction
AlixPartners
AI & Data - Vice President, Paris
Generative AI and Large Language Models (LLMs)
Taught InS7 · #82How LLMs Are Made
Doctolib
Senior AI Engineer - Patient Health Platform, Paris
Establish strong baselines first (prompts, RAG, model selection) before reaching for custom ML.
Taught InS15 · #174Prompting vs RAG vs Fine-Tuning
Skills Checklist
The skills this role is hired for, each with the free lesson that teaches it.
- 01
Linear algebra and probability for AI
S5 · #58Probability and Statistics
- 02
Gradient descent and backpropagation
S7 · #73A Neural Network From Scratch
What is a Neural Network? - 03
Deep learning with PyTorch
S7 · #75Training Loop in PyTorch
What is Deep Learning? - 04
Transformers and attention
S7 · #80Transformers Explained
What is a Transformer? - 05
Pretraining, fine-tuning and RLHF
S7 · #82How LLMs Are Made
What is Fine-Tuning? - 06
Building a GPT from scratch
S15 · #183Build an LLM From Scratch 1
What is a Large Language Model? - 07
Reading research papers
S15 · #185Reading AI Papers
- 08
Training and evaluating a language model
S15 · #184Build an LLM From Scratch 2
What is LLM Evaluation?
Your Path
Every lesson of this path, season by season. Published lessons open their page.
Season 0: Start Here
8 Lessons- S0 · #1Learn AI to Get Hired: The Full MapLive
- S0 · #2What AI Really Is: AI vs Machine Learning vs LLMsLive
- S0 · #3AI Jobs Explained: Which Role Fits You?Live
- S0 · #4Why Code Isn't the Hard Part AnymoreLive
- S0 · #5Escape Tutorial Hell: Learn by BuildingLive
- S0 · #6AI Assistants for Learners: Catch the MistakesLive
- S0 · #7Prompting from Zero: Get Better Answers
- S0 · #8Tokens, Context Windows and Model Families
- S1 · #9How a Computer Works
- S1 · #10Operating Systems
- S1 · #11Files, Folders, Paths, Formats
- S1 · #12The Command Line From Zero
- S1 · #13Linux Essentials
- S1 · #14How the Internet Works
- S1 · #15HTTP From Zero
- S1 · #16JSON and YAML
- S1 · #17Your Code Editor
Season 2: Python
17 Lessons- S2 · #18Install Python, First Script
- S2 · #19Variables, Types, Operators
- S2 · #20Strings and f-strings
- S2 · #21Lists, Dicts, Tuples, Sets
- S2 · #22Conditionals and Loops
- S2 · #23Functions
- S2 · #24File I/O and Parsing
- S2 · #25Errors and Exceptions
- S2 · #26Modules, Packages, Dependencies
- S2 · #27Classes and Objects
- S2 · #28Type Hints and Pydantic
- S2 · #29Reading Code You Didn't Write
- S2 · #30Calling Web APIs in Python
- S2 · #31Async Python
- S2 · #32High-Concurrency Python
- S2 · #33Testing With pytest
- S2 · #34Mini-Project: A CLI Tool
Season 3: Git and GitHub
6 Lessons- S3 · #35Version Control and Git
- S3 · #36Branches and Merges
- S3 · #37GitHub
- S3 · #38Pull Requests and Code Review
- S3 · #39CI/CD Introduction
- S3 · #40A Portfolio-Grade Repo
Season 4: Data Foundations
15 Lessons- S4 · #41What Data Is
- S4 · #42SQL 1
- S4 · #43SQL 2
- S4 · #44SQL 3
- S4 · #45Database Design and PostgreSQL
- S4 · #46Python Meets SQL
- S4 · #47NumPy
- S4 · #48pandas
- S4 · #49Data Visualization
- S4 · #50Business Intelligence and Data Storytelling
- S4 · #51NoSQL
- S4 · #52Data Engineering Basics
- S4 · #53The Modern Data Stack
- S4 · #54Data Governance and Master Data Management
- S4 · #55Ingesting Messy Documents
Season 5: Math for AI
6 Lessons- S5 · #56Vectors and Matrices
- S5 · #57Dot Product and Cosine Similarity
- S5 · #58Probability and Statistics
- S5 · #59Derivatives and Gradients
- S5 · #60Gradient Descent
- S5 · #61Loss, Softmax, Cross-Entropy
Season 6: Classical Machine Learning
11 Lessons- S6 · #62What Machine Learning Is
- S6 · #63First Model With scikit-learn
- S6 · #64Classification
- S6 · #65Evaluating Models
- S6 · #66Features and Pipelines
- S6 · #67Trees and Boosting
- S6 · #68Clustering and Dimensionality Reduction
- S6 · #69Time Series and Anomaly Detection
- S6 · #70Optimization and Simulation
- S6 · #71Experiment Tracking
- S6 · #72Project: Predictive Maintenance
Season 7: Deep Learning and Transformers
10 Lessons- S7 · #73A Neural Network From Scratch
- S7 · #74PyTorch Fundamentals
- S7 · #75Training Loop in PyTorch
- S7 · #76TensorFlow and Keras
- S7 · #77CNNs and Computer Vision
- S7 · #78NLP Foundations
- S7 · #79From RNNs to Attention
- S7 · #80Transformers Explained
- S7 · #81Tokenization Deep Dive
- S7 · #82How LLMs Are Made
- S15 · #174Prompting vs RAG vs Fine-Tuning
- S15 · #175Preparing Fine-Tuning Data
- S15 · #176LoRA and QLoRA
- S15 · #177Preference Tuning
- S15 · #178Model Optimization
- S15 · #179Small Models and Edge AI
- S15 · #180Training at Scale
- S15 · #181The Advanced ML Toolbox
- S15 · #182Scientific AI
- S15 · #183Build an LLM From Scratch 1
- S15 · #184Build an LLM From Scratch 2
- S15 · #185Reading AI Papers
Season 17: Getting Hired
9 Lessons- S17 · #211Choose Your AI Role: Decode Any Job Ad
- S17 · #217Your AI CV
- S17 · #218Build in Public
- S17 · #219Networking
- S17 · #220Certifications Worth Your Time
- S17 · #221System Design Interviews
- S17 · #222Coding Interviews in the AI Era
- S17 · #223Case Study and Business Case Interviews
- S17 · #224Behavioral Interviews and Negotiation
The Books Behind This Path
The lessons on this path were researched from these books. Read them to go deeper:

Understanding Deep Learning
Simon J. D. Prince, MIT Press, 2023
Seasons 5 to 7, 14 and 15
Free Official Copy
Probabilistic Machine Learning: An Introduction
Kevin P. Murphy, MIT Press, 2022
Seasons 5 and 6
Free Official Copy


Mathematics for Machine Learning
Marc Peter Deisenroth, A. Aldo Faisal and Cheng Soon Ong, Cambridge University Press, 2020
Seasons 5, 7 and 15
Free Official Copy
Related Paths
The paths that share the most lessons with this one.
Never Miss a Lesson
Subscribe on YouTube, or get an email the day each new lesson goes live.