Apply now
Master'sAutonomous, ADAS & Software-Defined Vehicle Applications open

Master’s in Reinforcement and Imitation Learning Engineering for Driving

Online master's programme of 1,900 hours in reinforcement and imitation learning engineering for driving, structured in 12 modules covering deep networks for control, imitation learning, inverse RL, deep and model-based RL, offline RL, simulation and sim-to-real transfer, learning safety, foundation models and deployment, with an assigned technical tutor.

Duration1900 h
ModeOnline
LanguageEnglish
Start datesSet by the faculty
AwardSEIUM Institute diploma
Enrol now View syllabus Ask an advisor
Virtual lab · try it before you enrol

Take the controls.

Run a real engineering trade-off from this field. Move the sliders, watch the model react in real time and beat the challenge.

SIM · LIVE ●
Learning outcomes

What you will be able to do

01

Apply deep networks to the control of automated driving.

02

Train policies through imitation learning.

03

Infer reward functions with inverse learning.

04

Implement deep reinforcement learning algorithms.

05

Transfer policies from simulation to real environments.

06

Assess the safety of learned policies.

07

Develop a learning-based driving project.

Profile

Who it's for

  • Machine learning engineers for driving.
  • Reinforcement learning researchers.
  • Learned planning engineers.
  • Simulation engineers.
  • MLOps engineers.
  • Driving data scientists.
  • Engineering or science graduates looking to enter the automotive sector.
Standards & tools

What you'll work with

  • ISO 21448 (safety of the intended functionality, SOTIF)
  • ISO 26262 (functional safety of road vehicles)
  • ISO/PAS 8800 (safety and artificial intelligence in road vehicles)
  • Regulation (EU) 2024/1689 (Artificial Intelligence Act)
  • ISO/IEC 22989 (artificial intelligence concepts and terminology)
  • ISO/IEC 23894 (AI risk management)
  • UL 4600 (safety of autonomous products)
  • SAE J3016 (levels of driving automation)
  • PyTorch
  • CARLA
  • highway-env
  • Stable-Baselines3
  • Weights & Biases
Syllabus

12 modules · 1900 h

  1. 1.1Learning paradigms in automated driving
  2. 1.2End-to-end versus modular driving
  3. 1.3Human driving data as supervision
  4. 1.4Simulation as a source of experience
  5. 1.5The distribution shift problem
  6. 1.6Safety of learned systems
  7. 1.7Industrial cases of learned driving
  8. 1.8Open research challenges
  9. 1.9Comparison of published approaches
  10. 1.10Case study: analysis of an end-to-end system
  1. 2.1Convolutional networks for visual input
  2. 2.2Recurrent networks for time sequences
  3. 2.3Transformers and attention mechanisms
  4. 2.4Bird's-eye-view representations
  5. 2.5Distributed GPU training
  6. 2.6Regularisation and augmentation techniques
  7. 2.7Optimisers and learning-rate scheduling
  8. 2.8PyTorch and JAX frameworks
  9. 2.9Experiment tracking and reproducibility
  10. 2.10Lab: control network in PyTorch
  1. 3.1Supervised behaviour cloning
  2. 3.2Covariate shift in imitation
  3. 3.3The DAgger algorithm and data aggregation
  4. 3.4Imitation conditioned on navigation commands
  5. 3.5Imitation with multimodal outputs
  6. 3.6ChauffeurNet and perturbation of demonstrations
  7. 3.7Imitation of privileged expert planners
  8. 3.8Driving demonstration datasets
  9. 3.9Closed-loop policy evaluation
  10. 3.10Practical: behaviour cloning in CARLA
  1. 4.1Formulation of the IRL problem
  2. 4.2Maximum entropy IRL
  3. 4.3Generative adversarial imitation learning (GAIL)
  4. 4.4Learning cost functions
  5. 4.5Learning from human preferences
  6. 4.6Reward ambiguity and identifiability
  7. 4.7Scalability to continuous spaces
  8. 4.8Modelling driving style
  9. 4.9Evaluation of learned rewards
  10. 4.10Workshop: IRL of motorway behaviour
  1. 5.1Markov decision processes
  2. 5.2Q-learning and DQN networks
  3. 5.3Policy gradient methods
  4. 5.4PPO and SAC algorithms
  5. 5.5Continuous actor-critic architectures
  6. 5.6Efficient exploration strategies
  7. 5.7Reward design for driving
  8. 5.8Simulation environments for RL
  9. 5.9Training stability and reproducibility
  10. 5.10Lab: PPO in highway-env
  1. 6.1Learned world models
  2. 6.2The Dreamer algorithm and latent space
  3. 6.3Planning with MuZero
  4. 6.4Planning with learned models
  5. 6.5Epistemic model uncertainty
  6. 6.6Sample efficiency versus model-free
  7. 6.7World models for driving
  8. 6.8Learned neural simulators
  9. 6.9Evaluation of model quality
  10. 6.10Practical: world model in CARLA
  1. 7.1Formulation of offline reinforcement learning
  2. 7.2CQL and IQL algorithms
  3. 7.3Use of fleet data
  4. 7.4Off-policy evaluation
  5. 7.5Distribution bias in logged data
  6. 7.6Combining offline RL and imitation
  7. 7.7Scaling with large data volumes
  8. 7.8Dataset curation and filtering
  9. 7.9Limitations of offline learning
  10. 7.10Workshop: CQL with driving data
  1. 8.1CARLA and MetaDrive simulators
  2. 8.2Visual and dynamic domain randomisation
  3. 8.3Transfer from simulation to reality
  4. 8.4Photorealistic neural simulation
  5. 8.5Generation of adversarial scenarios
  6. 8.6Progressive curriculum learning
  7. 8.7Multi-agent traffic environments
  8. 8.8Massive scaling of simulation
  9. 8.9Validation of the transfer
  10. 8.10Lab: domain randomisation in CARLA
  1. 9.1Safe reinforcement learning
  2. 9.2Constraints during learning
  3. 9.3Runtime safety shields
  4. 9.4Formal verification of neural networks
  5. 9.5Robustness against adversarial attacks
  6. 9.6Out-of-distribution input detection
  7. 9.7Uncertainty estimation in policies
  8. 9.8Runtime monitoring
  9. 9.9Safety argumentation for AI
  10. 9.10Practical: safety shield over PPO
  1. 10.1Vision-language models
  2. 10.2Language models for road reasoning
  3. 10.3Generative driving video models
  4. 10.4Tokenisation of actions and trajectories
  5. 10.5Explainable reasoning of decisions
  6. 10.6Large-scale multimodal data
  7. 10.7Computational cost and latency
  8. 10.8Risks of hallucination and bias
  9. 10.9Current research trends
  10. 10.10Workshop: VLM for scene description
  1. 11.1Closed-loop driving metrics
  2. 11.2CARLA Leaderboard benchmark
  3. 11.3nuPlan and Bench2Drive benchmarks
  4. 11.4Distillation of large models
  5. 11.5Quantisation for embedded hardware
  6. 11.6Deployment on automotive platforms
  7. 11.7Shadow-mode validation
  8. 11.8Regulatory requirements for AI
  9. 11.9Accountability and traceability of decisions
  10. 11.10Case study: deployment in shadow mode
  1. 12.1Definition of the problem and ODD
  2. 12.2Selection of the learning paradigm
  3. 12.3Data preparation and curation
  4. 12.4Policy training
  5. 12.5Integration of safety mechanisms
  6. 12.6Closed-loop simulation evaluation
  7. 12.7Analysis of results and failures
  8. 12.8Conclusions and future work
  9. 12.9Master's Thesis report
  10. 12.10Master's Thesis: defence before a panel
Methodology, software & labs

How you'll learn

The programme is 100% online on the SEIUM campus, with downloadable material, recorded classes and an assigned technical tutor. There are no fixed timetables: you progress at your own pace within the intake, so you can combine it with your job.

Each module combines theory, guided reading of the original standard and applied exercises. The working software is PyTorch, CARLA, highway-env, Stable-Baselines3 and Weights & Biases, in free, academic or evaluation versions. Testing is studied through real procedures for policy training in simulation, evaluation on the CARLA Leaderboard and shadow-mode policy validation. Assessment is continuous by module, plus the final Master's Thesis.

Standards covered: ISO 21448, ISO 26262, ISO/PAS 8800, Regulation (EU) 2024/1689 (AI Act), ISO/IEC 22989, ISO/IEC 23894, UL 4600, SAE J3016.

Assigned technical tutor:every module ends with a deliverable that your tutor marks individually.
Standards-based learning:guided reading of ISO 21448, ISO 26262 and ISO/PAS 8800 applied to real cases, not as a theoretical annex.
Industry software:PyTorch, CARLA, highway-env, Stable-Baselines3 and Weights & Biases, in free, academic or evaluation versions.
Master's Thesis with defence:A learned driving policy for a defined ODD, from data curation and training to safety mechanisms and closed-loop evaluation, defended before a panel.
Flexible format:100% online, no fixed timetables, compatible with full-time work.
Capstone project

Build it end to end

Your final project follows the real engineering workflow, step by step, reviewed by your tutor.

Problem and paradigm
Define the problem and ODD and select the learning paradigm.
Data preparation
Prepare and curate the training data.
Policy training
Train the policy with imitation or reinforcement learning.
Safety mechanisms
Integrate safety shields and uncertainty monitoring.
Master's Thesis
Evaluate the policy in closed-loop simulation and analyse failures, defended before a panel.
Career opportunities

Where this takes you

Machine Learning Engineer for DrivingReinforcement Learning ResearcherLearned Planning EngineerSimulation EngineerMLOps EngineerDriving Data Scientist
Admissions & scholarships

Requirements

  • University degree in engineering, science or architecture, or equivalent demonstrable professional experience.
  • Internet connection and a computer capable of running the programme's software.
  • English sufficient to read technical standards.

Ready to start?

Send your application today. Our admissions team replies within 72 hours with your study plan and funding options.

Apply now Scholarships
FAQ

Frequently asked questions

The programme totals 1,900 hours over 12–14 months. You set your own pace within each intake, and your tutor helps you plan the workload around your job.

You work with PyTorch, CARLA, highway-env, Stable-Baselines3 and Weights & Biases, using free, academic or evaluation versions. Your tutor indicates the free alternative wherever one exists.

A SEIUM Institute diploma, awarded on passing the continuous assessment and the Master's Thesis. It is SEIUM's own qualification, not an official state degree; its recognition depends on the programme completed and the regulations of each country.

Yes. Modules passed in a course or postgraduate diploma in the same area are recognised in the higher programme. Ask about the recognition policy when you enrol.

A university degree in engineering, science or architecture (or equivalent professional experience), a computer able to run the software, and English sufficient to read technical standards.

Start dates are set by SEIUM's faculty for each intake. Our admissions team will confirm the next available date when you apply.

Keep exploring

Related programmes

See all in Autonomous, ADAS & Software-Defined Vehicle
Master's9,970 $

Master’s in Perception & Sensors Engineering

Online master's programme of 1,900 hours in perception and sensors engineering. 12 modules covering perception for automated vehicles, cameras and computer vision, automotive radar, LiDAR, ultrasonic…

1900 hOnlineAutonomous, ADAS & Software-Defined Vehicle
Master's9,970 $

Master’s in Functional Safety, Cyber and Compliance Engineering

Online master's programme of 1,900 hours in functional safety, cybersecurity and compliance engineering, structured in 12 modules covering ISO 26262, SOTIF under ISO 21448, cybersecurity under…

1900 hOnlineAutonomous, ADAS & Software-Defined Vehicle
Master's11,090 $

Master’s in Robust Perception in Adverse Conditions Engineering

Online master's programme of 1,900 hours in robust perception engineering for adverse conditions. 12 modules covering adverse conditions and ODD, the physics of sensor degradation, cameras,…

1900 hOnlineAutonomous, ADAS & Software-Defined Vehicle
Master’s in Reinforcement and Imitation Learning Engineering for DrivingEnrol
Related programmes

More in Autonomous, ADAS & Software-Defined Vehicle

Master’s in Optimal Control and MPC Engineering for Autonomous Vehicles — Constraints, Robustness and Real TimeView programme →Master’s in C-V2X / ITS-G5 and Cooperative Perception EngineeringView programme →Master’s in Robust Perception in Adverse Conditions EngineeringView programme →Master’s in Diagnostics, Telemetry and Edge Analytics EngineeringView programme →Master’s in MLOps Engineering for AD/ADASView programme →Master’s in SDV/ADAS Product Management EngineeringView programme →

All Autonomous, ADAS & Software-Defined Vehicle programmes → · How to apply · Fees & financing · Scholarships

Scroll to Top
SEIUM – Institute of Advanced Engineering
Powered by  GDPR Cookie Compliance
Privacy Overview

This website uses cookies so that we can provide you with the best user experience possible. Cookie information is stored in your browser and performs functions such as recognising you when you return to our website and helping our team to understand which sections of the website you find most interesting and useful.