Editorial standards
Only what can be verified.
RobotGuide is a knowledge map for people who want to enter robotics. No courses for sale, no job guarantees, no "zero experience to six figures" promises. Data comes with sources; judgments are labeled as judgments; anything not measured is never written up as measured.
What this site solves
Entry-level robotics information is extremely fragmented: half university curricula, half vendor keynotes. The middle part — what to learn in what order, how deep is enough, what interviews actually ask — is what RobotGuide systematically covers.
Every one of the 74 topics states what to learn, the depth each of the four career tracks requires, how to verify you have learned enough, and the common mistakes. On top of that: lab benches you can actually drag parameters in, so "what happens if I change this" becomes visible; and architecture teardowns of six representative robotics companies, so you can see how the topics fit into real machines.
Where the content comes from
Job postings
Original JD text from public listings: salary ranges, degree requirements, responsibilities, skill demands.
Official docs & code
Vendor documentation and open-source repositories, read directly rather than through press coverage.
Public curricula
The knowledge systems of public textbooks and courses: Modern Robotics, Probabilistic Robotics, the official ROS 2 documentation, and similar.
Wherever a statement is inference rather than fact, it says so explicitly. Wherever something has not been measured, it is not presented as measured.
Principles
No hype
"A million-person talent gap" is everywhere. The actual shape of the gap is narrow at the top, wide at the bottom — and the bottom is field-service work. This site writes the real distribution.
No "full-stack"
Robotics spans five disciplines; true full-stack people barely exist. Claiming "full-stack" in an interview is self-incrimination.
Acceptance criteria, not hour anxiety
"Can you explain X clearly" is a far better test than "how many hours you logged".
Experiments are not toys
Every lab bench maps to real interview questions. After playing, you should be able to quote numbers, not just "I dragged the slider".
Known limitations
Hour estimates are engineering-scale figures for "zero to interview-ready"; individual variance is large. Project cost estimates are based on Chinese retail parts prices and float with the market. Salary data is concentrated in Shanghai listings; discount for other cities. Lab models are simplified: magnitudes match real systems, but they are not a substitute for measurements.
If you find an error or an unclear topic, write in — this site keeps changing. Contact: daibinhua888@163.com (Chinese community presence: Xiaohongshu "McKay Dai").
Two editions, one process
RobotGuide publishes in two editions sharing one editorial process, one source base, and one author: the English edition at robotguide.tech and the Chinese edition at robotguide.dev. Facts and data are shared across both; opinions are the author's own and labeled as such. The Chinese edition currently carries the complete library; English pages are added batch by batch.