Think Autonomous
⚠️ If your team keeps saying "We want to do End-To-End, but we're blocked"... this bundle has been created FOR YOU!

How to become the trusted "Go-To" engineer on End-To-End in your company

Everyone sells End-To-End as "data in, trajectory out". Here's what it really takes to make it work in a self-driving car.

End-To-End Revolution + Video Perception Get straight to End-To-End →Until Thursday, October 1 · 595€

Offer closes in ...

Have you been pressured by your boss to learn everything about End-To-End "for yesterday"... and still don't know if you have all the keys?

Then these are a few of the ideas you'll find inside two of my courses, bundled together until Thursday, October 1: End-To-End Revolution, your AV 2.0 core, and Video Perception, your entry into AV 3.0.

You can read more about who's teaching them in the box on the right, if you want.

Otherwise, here's a small taste of what's inside:

End-To-End Revolution

Course 1 · AV 2.0

End-To-End Revolution

A revolution is happening. End-To-End Learning has emerged, and hundreds of startups are converting to learned approaches. Become the engineer who knows how to build it, and how to explain it.

  • Advanced
  • 2 projects: E2E driving + visualization
Opening the black box: watching the attention shift inside an End-To-End model
The End-To-End project: camera + LiDAR fusion, depth, attention and trajectory

Quick break. Why these two courses together?

Because AV 3.0 is built on top of AV 2.0. When Alibaba's Qwen released its self-driving model, the core was a classic AV 2.0 model: HydraNets, Bird Eye View, Diffusion Planning. The AV 3.0 part was what came on top: a VLM "pre-encoder", and a training mixing Imitation and Reinforcement Learning.

And every one of these models works on video. Here's the second half:

Video Perception

Course 2 · AV 3.0

Video Perception

Get to the most advanced part of Computer Vision: master video processing, reasoning, and the skills AV 3.0 models are built on.

  • Advanced
  • Mini-project on real Waymo footage
Video QA: asking questions to a driving scene
The AV 3.0 Autonomy Flow

What engineers say about these two courses

Straight from my inbox and the course comments.

End-To-End Revolution

"By far the most interesting and in depth course"

I have been doing your end to end course for a while now, I have to say it, it is by far the most interesting and in depth course by far, the amount and depth of information has been absolutely humbling, so many good references and directions to pursue from here for me.

Being in an OEM, I hear a lot about foundational and end to end models being the future, not just in terms of predicting way points but also in other modular ways. This course has been an eye-opener to end to end tech for me.

Wonderful course, a lot to be learnt in there in the coming years. Absolutely worth the price!!

VP
VPName withheld pending authorization· email, September 2026
VP's email about End-To-End Revolution
Video Perception

"I'll say it here: wowzie, Video Perception is the best course"

I made a couple of comments in the course so I will also say this here: wowzie video perception is the best course and not just because it's recent.

Couldn't have asked for a better trio: optical flow, transformers on unique datasets, AV 3.0

Michael Buyer
Michael BuyerVideo Perception student · course community
Comments from Video Perception students: Mario Bergeron, Tiziano Cianti, Michael Buyer, Enrique Israel Hernandez

Become the End-To-End engineer your team turns to

The bundle closes Thursday, October 1 at 23:59 CEST.

0days
00hours
00minutes
00seconds

YES, Jeremy! I want to become the End-To-End engineer my team turns to. Give me instant access to End-To-End Revolution and Video Perception for 595€ instead of 844€.

I understand this price ends Thursday, October 1 at 23:59 CEST.

This bundle is now closed.

The offer ended on Thursday, October 1 at 23:59 CEST.

Frequently Asked Questions

Who is this bundle for and not for?

It's for engineers already in, or close to, the AI and self-driving car industry who want to be in control of the latest technologies. If you have never heard about End-To-End before, you're probably too early. If you're looking for a job, End-To-End Revolution is primarily built for people already in the industry, and you're better off learning Kalman Filters and Stereo Vision first.

What are the prerequisites?

Coding in Python, Deep Learning basics with PyTorch (Backpropagation, CNNs, MLPs), High-School Maths, Intermediate Computer Vision (Depth Estimation, Segmentation, Object Detection), and Modular Self-Driving Car Theory. Transformers, Bird-Eye View and HydraNets are optional.

What is the format like?

Self-study online courses, with videos, articles, drawings, paper analysis, code and projects. End-To-End Revolution takes ~5-10 hours, Video Perception ~5-7 hours.

Do I get support if I'm stuck?

Yes, the courses are hosted on the Think Autonomous 2.0 platform, optimized for collaboration, chat, support, answers, and community learning.