A Group of People, a New Journey
Through this journey, we hope to contribute to robotics in China and around the world!

The Beginning
October 2022. I was still doing my master's, but I was already interning at my advisor's company. My advisors treated us well, giving us the chance to work on cutting-edge projects, and the stipend was quite good.
As a graduate student, this kind of life was already comfortable, so I am truly grateful to my advisors.
But for someone like me, who had idolized Musk since freshman year, a stable job was not my ideal. So I chose to leave and strike out on my own.
Some friends worried about me, afraid I wouldn't get my degree, but I cared more about doing things that would help me grow faster.
The Studio: ARSoft
In November, after leaving the company, I had no income, but as a student I was used to living frugally. I rented a dorm room in the suburbs, bought some cheap desks, chairs, and monitors, and started working.
After a while, some good friends were willing to join me, working part-time on our project.
In March 2023, after months of brainstorming, research, and experimentation, we got our first client.
At the same time, luckily, a close classmate was willing to come work full-time without asking for any compensation.

That's how we started our first official job—providing control algorithms for the client's robot dog.
Here, I need to thank Mr. Wang and Mr. Chen from Unitree for lending us a robot dog free of charge to validate our algorithms. Their robot dogs are excellent.
In May, after we essentially finished the algorithm, we went to the client's site and spent two weeks there, getting their robot dog to walk quite well. That's when we earned our first revenue.
Meanwhile, based on our control framework, we made something fun on the side—getting a quadruped robot to walk on two legs.
This was the first of its kind in China at the time, and we used a relatively complex model-based approach. Even internationally, I had only seen one bipedal walking result using a learning-based method.
This work caused a bit of a stir in the legged robotics community, and that's when people in the field started to know me.
Actually, this project took only 2 days from start to physical implementation.
Three months earlier, when I first discovered how flexible OCS2 was, I was amazed. So in my free time, I tried it on the real robot, and it was indeed impressive.
I have to admire ETH's work and their contribution to the open-source community. I also need to thank Liao Lao for the application framework he built on hardware—it's really good, and we still use it now.
The Company: BridgeDP
At the end of May, the first collaboration went well, and the client chose to continue. With that modest first income, we registered a company and moved into the city, planning to hire a few people and try to survive another six months.
But hiring for a startup is full of pitfalls, and I stepped into quite a few. Money ran out fast, and I had to borrow from family to keep the company afloat.
In July, a number of people left the company, leaving only those with hope and dreams who chose to stay through thick and thin. It was a low point, but we kept developing. Fortunately, a startup doesn't need many people anyway. We produced some new work, as shown in the video below.
Around that time, we received an interview invitation from MiraclePlus and eventually got an offer.
In today's market, getting funding for a startup is hard, especially for unknown young people like us.
It's admirable that MiraclePlus continues to help young people today. Of course, young people with nothing who still recklessly choose to start a company today are probably as crazy as I am.
Dr. Lu Qi in the north and Professor Li Zexiang in the south are probably the two mentors in China who support young entrepreneurs the most. Thank you both!
Later, more talented people joined the company, all more capable than me in their own ways. Clients also started reaching out, and we took on more projects, getting hands-on with various legged robots.

Gradually, we became a company specializing in legged robot control systems—something I originally wanted to do. There are always skeptics, but that doesn't stop us from experimenting.
Today, we are officially releasing an open-source bipedal robot control algorithm and framework. The code will be open-sourced at the link below after cleanup:
https://github.com/bridgedp/hunter_bipedal_control
We hope to turn legged robot controllers into a module that can be adapted to all kinds of legged robots with one click—quadrupedal, bipedal, wheel-legged, and more. And it must be cheap enough for everyone to afford.
Only then can the supply chain for legged robots mature, downstream integrated applications flourish, and general-purpose legged robots become possible.
Finally, we welcome friends who are interested and don't mind a modest salary to join us; interns are welcome too.
The Future
Whether the legged robot controller business can succeed is still unclear, and the technology isn't fully ready. There's certainly a long road ahead, full of difficulties, but since I chose to start a company, it was never to take it easy.
Youth may be an advantage for the young. Taking on a challenge that helps me grow, connects me more deeply with the world, and lets me meet interesting friends—isn't that a kind of happiness?
Of course, as my own daunting challenge, I need to ensure the happiness and well-being of my colleagues, so figuring out how to make money is my job.
If you, esteemed reader, have any good advice, feel free to reach out (WeChat ID: wx-of-syx).
Last of all, I hope we can overcome every obstacle and, through this journey, contribute to robotics in China and around the world!





界面新闻
深视新闻
深圳特区报
财新
36氪
钛媒体
中国新闻周刊
蓝鲸新闻
Z Finance
投资家网
雷峰网
亿欧
环球网
南方日报
极客公园



