Shopify

Product Lead

Senior Product Manager

Angel Investor

Active:
* Aquaeye by Vodasafe
* Evalyze
* EverBond Interface Technologies
* HDAX Therapeutics
* Healthy Hippo
* Intellijoint Surgical
* KaleMart24
* Kenota Health
* MesoMat
* Mirsee Robotics
* ORA Graphene Audio
* OrderEase
* Pontosense
* Quip Medical
* Steadiwear
* ToroMatcha
* Trampoline.ai
* Vena Medical
* VoxCell BioInnovation

Completed:
* Little & Lively - Dead

Golden Triangle Angel Network (GTAN)

Vice Chair Board of Directors

GTAN is the #1 Seed Stage Private Investor in Canada.

Board Member

Grobo Inc.

Founder and CEO

Entrepreneur, Product Manager, Advisor & Angel Investor with experience building strong products and companies.

Bjorn Dawson

Tianyu Xu

TYAI

Generative AI is transforming the way we work. Understanding its capabilities, ensuring its safe use, and knowing how to leverage it to solve business challenges are critical. We empower people with generative AI skills and advocate human-AI synergy.

Key offerings:

↳ Tailored AI workshops for companies  
↳ Generative AI training (public)  
↳ Consulting

Founder | Generative AI Educator

Future of Marketing Institute

Member of the Advisory Board

Twitter

Client Account Manager (Large Client Solutions)

Global Client Account Manager

Fonterra

Global Store Analytics Lead

Research Analyst (Marketing Insights & Analytics), APAC

We Are Social

Senior Research & Insights Executive

I help people acquire future-proof generative AI skills

I help people work smarter with AI and stay ahead. I co-created two top-selling AI image creation courses and regularly speak publicly on AI trends, AI videos, and creative AI tools.

𝗥𝗲𝘀𝗼𝘂𝗿𝗰𝗲𝘀

↳ Publication: Will ChatGPT Take My Job [Amazon] 
↳ Course: ChatGPT Image Generation Master Course [Udemy Business]
↳ LinkedIn: @Tianyu Xu – dedicated to sharing knowledge
↳ TikTok: @MrAICat – focused on entertainment
↳ eBooks & recordings: tianyuxu.gumroad.com
↳ Portfolio: tianyuxu.com

𝗦𝗲𝗿𝘃𝗶𝗰𝗲𝘀

↳ Generative AI training for companies  
↳ Speaking engagement
↳ AI video production
↳ Consulting

𝗟𝗲𝘁'𝘀 𝗰𝗼𝗻𝗻𝗲𝗰𝘁

Whether you’re seeking to empower your team with generative AI or are interested in exploring AI-powered creativity, I’m here to help. Let’s discuss AI and find new possibilities!


Past and current brand partnerships / CPP: Luma AI, Minimax Hailuo, OpenArt, Leonardo, Manus, and more.


<More Context>

Tianyu brings over 2,000 hours of hands-on generative AI experience, crafting solutions that boost efficiency and productivity for individuals and teams. His programs enable natural adoption of AI tools, sparking creativity, streamlining workflows, and offering a competitive edge. An expert in accessible generative AI, Tianyu enhances human potential through creativity and efficiency.

IMPACT

- AI Education: Tianyu co-created two bestselling AI image generation courses and launched multiple live AI video courses, empowering thousands of professionals in image and video generation. He co-authored a ChatGPT book with practical strategies for leaders.

- Content Creation: He’s produced tens of thousands of AI-generated images and videos, sharing insights on social media to over 150k social media followers with millions of views.

- Speaking Engagement: Tianyu presents at industry events in Asia and North American institutions, focusing on AI for creativity and productivity, and collaborates with experts on topics like "Generative AI for Architects."

OFFERINGS

- Custom Workshops for Businesses: Team-focused sessions on tools like ChatGPT, Gemini, Midjourney and Luma to optimize workflows, e.g., accelerating marketing campaigns.
- Public Trainings for Professionals: Hands-on courses covering generative AI from basics to advanced, ideal for upskilling.
- Consulting: Strategic AI guidance and tailored plans for organizations to achieve their goals.

BENEFITS

Tianyu’s training leverages AI to automate routine tasks, freeing you or your team for high-impact work.

</More Context>

Founder

Generative AI is evolving at lightspeed. Understanding its capabilities, ensuring its safe use, and knowing how to leverage it to solve business challenges is critical. If you can enhance your business now, why wait? Here's how we can help:

1. Generative AI Consulting
Designed to help businesses unlock the potential of generative AI technologies and tools (ChatGPT, Midjourney, Bard, etc.). These include assessing AI opportunities, strategizing, and executing implementations that drive business innovation and efficiency.

2. Generative AI Workshop
Tailored to equip teams with the knowledge needed to effectively use generative AI tools, ensuring they can apply these skills from day one.

3. AI-Powered Research and Analytics
This offering provides businesses with timely insights and assistance in identifying growth opportunities, facilitating informed, data-driven decision-making.

1 attempt, 2 prompts, 13 videos, 10 minutes👇 
You just need a reliable AI video model that works.
[Recorded Course: LoRA for Architects https://lnkd.in/g3mcuGtS]

All videos here were generated using Kling 2.1 
start/end frames, in a Weavy workflow. 

It feels old-school, but I always advise people that 
what you need is not the latest, fancy AI model, 
but something that performs for your use case.

Sora 2 is great, Veo 3 is fascinating (I’m addicted to it), 
but if you want to easily bring images to life 
with almost no quality issues, 
Kling is the best choice.

If you don't have time to test all the models 
and you are an architect or designer, 
Kling is your best video partner.

In the workshop tomorrow, 
my colleague Ar. June Chow will take you through 
the most practical ways to train a LoRA with the styles 
you desire, and I will share my video best practices.

For those who have signed up, see you soon!

Sora 2 vs. Veo 3 vs. Wan 2.5 vs. Kling 2.5 🍿
Which AI do you think is the winner here?
A side-by-side comparison.
Text-to-video only.

I used the prompts from OpenAI's Sora 2 release, 
curious to see how they perform on the others.

It seems that Sora 2 brought OpenAI back to 
the top league in the video space.
But it's too early to tell. 

From the videos shared by other social media users, 
I've seen mixed results.

Some outstanding capabilities of Sora 2:

⭐ Anime
⭐ Sports scenes 
⭐ Multi-camera shots

But if you know how to prompt Veo 3, 
results like this aren't hard to achieve.

Don’t forget there’s another alternative: Wan 2.5.

There’s no single model that rules them all. 
Identify their strengths and use them. 
Hopefully, these models will stay in your toolkit 
and fit into the right place in your workflow.

While waiting for access, 
I look forward to learning from those who 
already have their hands on it.

To our friends in the US and Canada, 
please feel free to share your results 
and findings in the comments!

How to build a Nano Banana-powered 
AI image app in 10 minutes (and it’s free)
A step-by-step guide with my prompt:

You can vibe code in Google AI Studio now. 
It works similarly to other vibe coding apps 
but as easy and simple as Custom GPTs.

Step 1: Visit Google AI Studio -> Build 
Step 2: Select an AI model or tool 
Step 3: Describe your idea 

5 to 10 minutes later, the version is ready.
You can spend hours finetuning the app. 

I built this profile photo app over the weekend 👇
(Thanks, Stefan Kirschnick for joining the test!)

Here’s my simplified prompt to start with:

Build an AI profile photo app. Dark, 3-panel layout. 
Left panel: photo uploads. After upload, open 
a crop tool with 1:1, 16:9, and 9:16 aspect ratios. 
Center panel: 2x2 gallery of generated images. 
Right panel: 'Pro', 'Casual', and 'Quote' styles. 
'Quote' style shows the character next to the 
quote text in 4 different settings. 

A few things to note if you are new:

↳ Google AI models and tools only (of course)
↳ API is required except Nano Banana and Gemini 2.5
↳ Protect your API keys if you are using them.
↳ Make your app simple.

Most people I spoke to in the past year 
have never tried Google AI Studio or vibe coding, 
but now you can have both in one place.

It’s the best chance to build something useful 
and make it part of your AI learning journey.

Sora 2 (API) is here, but how good is it, really? 
An image-to-video test with Sora vs competitors👇 
[📣Workshop: LoRA for Architects luma.com/ohu4jzon]

I started with an image from Ar. June Chow, 
based on a LoRA she trained on a specific style.

If you've attended my previous AI video trainings, 
you'll know that I always advocate for 
image-to-video as the only practical use of AI video. 
And prompting with camera-related keywords 
can easily bring your image to life.

Here are my initial takeaways:

🟢 Sora 2 has high prompt adherence, 
making it useful for storytelling and education.

🟢 What’s unique: It recognizes time and numbers 
and is able to display 3-4 shots in one video.

🤔 The success rate of getting an ideal video is about 
the same as with Veo 3 but lower than with Kling.

🤔 The graphic quality of the videos is relatively low, 
even if you choose 1080p.

🤔 It’s relatively hard to achieve smooth 
and consistent camera motions.

In the video below, I generated the Sora videos 
on Higgsfield and the others on Flow and OpenArt.

Which AI model do you think is the winner here?

In my next post, I'll make another comparison 
for a different use case. 
Stay tuned!

Google's Veo 3.1 is available now! 
According to the official announcement, it has 

- richer native audio
- better cinematic styles
- reference to video 
- transitions between frames
- video extensions

In my tests today, 
I saw smoother lip-sync and motion, 
and a higher success rate.

I believe the Veo 3.1 start/end frame will be 
the real game-changer. 

Another highlight is the inpainting feature. 
(It was there earlier, but I didn't find it useful.)

Firstly, select an area of a video, 
then tell Veo 3.1 what you want to change, 
and it will edit the video for you. 

This creates more practical use cases.

As of now, 
the best place to use Veo is still Flow. 
If you have a Gemini account, 
you will have it.

I will test it more and share a more 
holistic review soon! 
Stay tuned!

What I like most about Sora: timing your shots.
Just tell AI the time, the shot, and the action.
[New Course] AI 101 for Architects luma.com/rzq3h77c

Prompt example in the video 👇 

In this way,  you can tell a consistent story. 
Instead of generating 5 different videos, 
you only need 1 video with 5 shots.

Limitations

12 seconds is the maximum duration for now. 
Random shots and low-quality graphics 
High chance of prompt rejection.

__________

About the New Course

The 3rd one of our "AI for architects" series.
Ar. June Chow and I will cover both design 
and non-design use cases of generative AI 
and help you build solid foundations.

No prior experience with AI is required.
Full recording will be provided.

taplio

The Battle

Bjorn Dawson

Tianyu Xu

Followers

Posts

Engagements

Likes

Performance ScoreTianyu Xu's LinkedIn Performance

Ready to grow your LinkedIn presence?

taplio

The Battle

Bjorn Dawson

Tianyu Xu

Followers

Posts

Engagements

Likes

Performance Score.css-goub6s{font-size:var(--chakra-fontSizes-md);font-weight:500;color:var(--chakra-colors-gray-600);display:block;margin-top:var(--chakra-space-1);}Tianyu Xu's LinkedIn Performance

Ready to grow your LinkedIn presence?

Performance ScoreTianyu Xu's LinkedIn Performance