Jean Lee

Most engineers are learning the wrong skills.
I heard the same thing over and over at RSAC.
The hiring market has already shifted.

I interviewed 5 tech leaders, and asked them one question: What are you actually looking for when you hire right now?

✅ Summary of lessons

1/ The shift happening in the tech job market
The job market is moving away from valuing specific tools and programming languages. Hiring is shifting toward adaptability, systems thinking, and the ability to navigate increasing complexity.

2/ What skills matter?
Technical depth alone is no longer enough. The most valuable skills now include problem framing, understanding systems, communication, and the ability to work effectively with AI.

3/ What matters for the job market?
Hiring managers are prioritizing mindset over experience with specific technologies. They are looking for candidates who can think critically, move fast, and understand how their work impacts the broader system.

4/ Is AI replacing jobs?
AI is automating parts of engineering work, especially repetitive tasks. However, it is also creating new roles and increasing demand for people who can manage, guide, and build on top of these systems.

5/ How to build the right skill
Building these skills requires going beyond tutorials and tools. Focus on understanding systems, studying real-world problems, and developing judgment through hands-on experimentation.

✅ Interview with:

1/ Mary Carmichael, Managing Director of Risk Advisory and a Member of ISACA

A risk and AI governance expert who has spent 15+ years advising organizations on how to navigate every major technology shift, from digital transformation to the age of agentic AI.

2/ Hom Bahmanyar, Global Enablement Officer at Ridge Security

A computer scientist turned global enablement leader who has watched the engineering skill stack evolve firsthand, and now shapes how security teams worldwide adapt to AI.

3/ Lydia Zhang, President and co-founder of Ridge Security

The co-founder and president of a cutting-edge offensive security company who oversees engineering teams and is actively redefining what she even looks for in a hire as AI rewrites the job.

4/ Shlomie Liberow, founder of Aisy and former head of Hacker R&D at HackerOne

The person who used to find the holes in the world's biggest companies at HackerOne, and now builds AI-powered security tools, someone who has seen both sides of what breaks and what holds.

5/ Jagannadh Mellacheruvu, Head of Product and Policy at Mirror Security

A product and policy leader at the frontier of AI data security, working on the encryption technology that protects what you send to AI, before it even gets there.

Watch the full episode 👇
I Asked 5 Tech Leaders What the Tech Job Market Really Wants Now https://youtu.be/HhZ50GpRO2I

5 days ago | [YT] | 188

Jean Lee

The only Claude Code cheat sheet you need!
Most engineers are still using AI tools wrong.
That’s why they’re falling behind.

Claude Code is a terminal-based agent that can:
→ write code
→ test it
→ and even commit changes
All from your repo.

𝗖𝗼𝗿𝗲 𝘄𝗼𝗿𝗸𝗳𝗹𝗼𝘄
Instead of typing prompts in a browser:
→ run claude in your project
→ plan
→ execute
→ verify
→ iterate inside your codebase
It’s closer to 𝗽𝗮𝗶𝗿 𝗽𝗿𝗼𝗴𝗿𝗮𝗺𝗺𝗶𝗻𝗴 𝘄𝗶𝘁𝗵 𝗮𝗻 𝗲𝗻𝗴𝗶𝗻𝗲𝗲𝗿 than using a chatbot.

𝗖𝗼𝗺𝗺𝗮𝗻𝗱𝘀 𝘄𝗼𝗿𝘁𝗵 𝗸𝗻𝗼𝘄𝗶𝗻𝗴
→ Shift + Tab → switch between Plan / Execute / Verify
→ /model → choose the right model for the task → /loop
→ automate repetitive workflows
→ /schedule → run tasks in the background
→ /compact → manage long context
Small commands. Big productivity difference.

𝗛𝗼𝘄 𝘁𝗼 𝗮𝗰𝘁𝘂𝗮𝗹𝗹𝘆 𝗴𝗲𝘁 𝗴𝗼𝗼𝗱 𝗼𝘂𝘁𝗽𝘂𝘁
Most people prompt like this:
“Fix this code”
That’s why results feel inconsistent.

Instead:
→ structure your prompts
→ separate code, requirements, and constraints
→ ask for step-by-step reasoning
→ define the level of effort (low vs max)
You’re not asking for code. You’re guiding a system.

𝗜𝗳 𝘆𝗼𝘂’𝗿𝗲 𝗯𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝘄𝗶𝘁𝗵 𝘁𝗵𝗲 𝗔𝗣𝗜
The setup is simple:
→ install anthropic
→ call client.messages.create()
→ pass clear system + user instructions
That’s enough to start building real applications.

𝗠𝗼𝗱𝗲𝗹 𝘀𝗲𝗹𝗲𝗰𝘁𝗶𝗼𝗻 (𝘀𝗶𝗺𝗽𝗹𝗲 𝘃𝗲𝗿𝘀𝗶𝗼𝗻)
→ Opus = complex reasoning, architecture
→ Sonnet = daily coding, iteration
→ Haiku = fast, lightweight tasks
Use the wrong model, and everything feels worse than it should.

💡 𝗧𝗵𝗲 𝘀𝗵𝗶𝗳𝘁
AI is no longer just answering questions.
It’s becoming part of your development workflow.
The engineers who adapt:
→ move faster
→ ship more
→ and learn quicker

If you’re still using AI like a chatbot, you’re underutilizing it.

📌 Watch >> 99% of Engineers Are Using AI Tools Wrong https://youtu.be/OvMoKZWtyKE

6 days ago | [YT] | 146

Jean Lee

AI Engineer vs. Data Scientist: What’s the real difference?
I analyzed hundreds of job postings to see which skills show up most often.

Here is what appeared in both roles:
1️⃣ Python
2️⃣ Machine Learning Fundamentals
3️⃣ AWS

Yes, there is overlap.But the work is not the same.

✅ Data Scientist
Data Scientists focus on models, data, and experimentation.
↳ Explore data and uncover patterns
↳ Design features and evaluate performance
↳ Run experiments and iterate
↳ Solve open-ended problems

Think:“What can we learn from this data, and what will happen next?”

✅ AI Engineer
AI Engineers focus on shipping AI into real products.
↳ Work with LLMs, APIs, RAG, and agent systems
↳ Deploy models into production environments
↳ Handle latency, cost, scaling, and reliability
↳ Integrate AI into existing software systems
↳ Monitor failures and improve system performance

Think:“How does this actually work in production?”

💡 The part most people miss
In real companies, these roles overlap a lot.
Titles vary more than the actual work.

🎯 The takeaway
Do not overthink the title.
Start with the role you can break into now.You can always switch later.
Careers are built through momentum, not perfect first choices.

Watch “I, Machine Learning, Data Science: Which is the Better Career” for a full breakdown
https://youtu.be/dadXZ7rrSfE

1 week ago | [YT] | 154

Jean Lee

I’ve never seen software engineers this anxious.
AI changed the mood completely.
A Meta is ramping up another round of layoffs.
People are losing sleep.

Everywhere you look:
→ AI will replace engineers
→ AI writes code now
→ Agents will automate software jobs
→ Maybe I should switch career

And honestly?
I understand the anxiety.
Even inside Big Tech, people feel the pressure.

At Meta, performance expectations are rising.
Layoffs are happening across the industry.
The bar keeps moving.

But here’s the thing most people miss:
AI is changing software engineering.
That part is real.

What’s NOT real is the idea that engineers suddenly become useless.
The engineers who adapt will still be valuable.
The skills I believe will matter even more now are:
→ communication
→ system design
→ project leadership
→ cross-team influence
→ judgment under ambiguity

Because building software was never just about typing code.

At RSAC, I interviewed multiple engineering leaders and asked:
“What skills will actually matter in the future?”

The answers were surprisingly consistent.
Not:
→ memorizing syntax
→ grinding LeetCode forever
→ trying to out-code AI

AI will remove some work. But it will also raise the leverage of strong engineers.

This transition period is messy.
A lot of people are scared.
You are not imagining it.
But panic is not a strategy.
Adaptation is.

🎥 I shared the full conversations and lessons from real hiring managers:
https://youtu.be/HhZ50GpRO2I

♻️ Repost if you’ve been feeling this pressure too
👋 Follow me, @Jean Lee, for grounded insights on AI and tech careers

1 week ago | [YT] | 89

Jean Lee

Most people wait for the perfect setup.
The right course.
The right certificate.
The right market conditions.

But learning AI does not start when everything is ready.

My YouTube videos got 6,406,950 views.
Most were filmed on my iPhone 14.

No studio.
No fancy lighting.
No expensive setup.

The lesson was not “gear does not matter.”

The lesson was:
starting beats over-preparing.

The economy will always feel shaky.
Your idea will always feel unfinished.
The market will never be perfectly timed.
The course will always feel incomplete.

I started filming before I felt ready.
I posted before I was sure.
I improved by doing.

That is how I built my channel.

Not with a perfect plan.
But by explaining one AI/career question at a time.

If you are waiting to start, start smaller.

Pick one topic you want to learn.
Build on what you learned.
Then repeat.

That is how confidence gets built.

Links to the most-watched videos:
👉 High-Paying Careers of the Future: What AI Won’t Take Over
965K views
https://youtu.be/U5XS7_FqL9k

👉 AI/ML Engineer path - The Harsh Truth
508K views
https://youtu.be/HR91Wrr8KH8

👉 AI Machine Learning Roadmap: Self Study AI!
493K views
https://youtu.be/nznFtfgP2ks

1 week ago | [YT] | 62

Jean Lee

I got invited to Google HQ last week.
Not as an engineer.
Not as a manager.
As media.

That sentence would have sounded ridiculous to an earlier version of me.

I went there to interview a Google VP about AI, Android, and what’s changing next.
But honestly, the bigger realization had nothing to do with Google.

It was this:
The internet changed the leverage equation completely.

You no longer need permission from traditional institutions to build a career, audience, or reputation.
A YouTube channel can become a media company.
A LinkedIn post can open doors to rooms you never imagined entering.
One idea can connect you to people who once felt completely inaccessible.

Most people still underestimate this.

They use these platforms passively.
The people who win use them intentionally.

The full interview drops next week.
And I think some of the insights are going to surprise people.

📌 See my latest coverage of the RSAC conference and interviews with 5 leaders in AI/security: https://youtu.be/HhZ50GpRO2I

1 week ago | [YT] | 178

Jean Lee

You don’t need to pay thousands of dollars to learn AI.
One of the most overlooked facts about the AI field is
The best resources are FREE.

While writing AI Engineering, Chip Huyen didn’t rely on bootcamps or expensive courses.
She went deep:
↳ 1,200+ reference links across papers, blog posts, tools, and real-world case studies
↳ 1,000+ tracked open-source GenAI GitHub repos
↳ Constant synthesis from practitioners actually shipping AI

The book itself became a map of the ecosystem, not another surface-level tutorial.
And the document she shared afterward?

It’s a curated list of the most useful resources for understanding every major area of modern AI:
👉 Stanford CS 321N
cs231n.github.io/
My favorite introductory course on neural networks.

👉 Apple's human interface guideline for designing ML applications
developer.apple.com/design/human-interface-guideli…

👉 Best Practices and Lessons Learned on Synthetic Data for Language Models (Liu et al., DeepMind 2024) arxiv.org/abs/2404.07503

👉 Challenges in evaluating AI systems by Anthropic
www.anthropic.com/news/evaluating-ai-systems

👉 16 Changes to the Way Enterprises Are Building and Buying Generative AI by Andreessen Horowitz
a16z.com/generative-ai-enterprise-2024/

👉 Retrieval-Augmented Generation for Large Language Models: A Survey
arxiv.org/abs/2312.10997

👉 Model Context Protocol (MCP): Landscape, Security Threats, and Future Research Directions
arxiv.org/abs/2503.23278

👉 Mastering LLM Techniques: Inference Optimization by Nvidia
developer.nvidia.com/blog/mastering-llm-techniques…

👉 Guidelines for Human-AI Interaction
www.microsoft.com/en-us/research/publication/guide…

This is how real engineers learn:
Not from “$5,000 bootcamp funnels”But from papers, repos, docs, and community knowledge.

The information is free.The discipline to go deep is the real cost.

If you need tips on how to affective read research papers and learn, watch, "STOP Taking Random AI Courses. Read These Instead"
https://youtu.be/KPhZkvPBiMk

1 week ago | [YT] | 43

Jean Lee

You don’t need to be smarter to get hired.
You need different skills.
Most CS majors never learn them.

I’ve interviewed hundreds of candidates.
The pattern is consistent.

Here are 4 mistakes keeping CS students unemployed:

1️⃣ Treating coding like a job
You can solve problems on your own.
But companies look for:
↳ Clear communication
↳ Structured thinking
↳ The ability to explain tradeoffs

Coding is expected.
Thinking out loud is evaluated.

📌Watch Should You Still Study Computer Science in the Age of AI?
https://youtu.be/n2hnbe_HBno

2️⃣ Not using AI tools
This one surprises people.
Even companies like Meta now allow AI tools in interviews.

Yet many students:
↳ Avoid them
↳ Or use them blindly
The goal is not to outsource thinking.
It is to use AI to move faster while still understanding why things work.

📌Watch 99% of Engineers Are Using AI Tools Wrong
https://youtu.be/OvMoKZWtyKE

3️⃣ Weak resume signal
Most resumes say:
“Built a project”
That tells me nothing.

Strong resumes show:
↳ What problem you solved
↳ What impact you had
↳ What tradeoffs you made
If your resume is vague, you get filtered out.

📌Get the free ATS friendly resume template:
exaltitude.substack.com/p/jeans-ultimate-ats-frien…

4️⃣ No real-world experience
No internships = no signal.
That doesn’t mean you’re stuck.
You can still build:
↳ End-to-end projects
↳ Systems with real users
↳ Anything that shows you can ship
Companies hire proof of work, not just coursework.

📌Watch Hiring Managers Don’t Care About Your Projects
https://youtu.be/Wb2smz8anas

💡 The part most people miss
It’s not about being “smart enough.”
It’s about showing:
→ how you think
→ what you’ve built
→ and how you operate in real systems

2 weeks ago | [YT] | 26

Jean Lee

The only Python roadmap you need
Python is now the most in-demand skill across software, data, and AI roles.
But you don’t need a $10K bootcamp to learn it.

—

𝗠𝘆 𝘀𝘁𝗼𝗿𝘆

I didn’t know Python when I started my software engineering career.
One day at work, I was asked to work on a project using Python.

I didn’t feel ready.I didn’t have a plan.

I just said:“I’ll figure it out.”
That moment forced me to learn fast.

𝗧𝗵𝗲 𝗮𝗰𝘁𝘂𝗮𝗹 𝗿𝗼𝗮𝗱𝗺𝗮𝗽
Most Python roadmaps are vague.
Here’s what actually matters:


1️⃣ Setup your environment
→ Install Python→ Pick an editor (VS Code, PyCharm, Jupyter)→ Learn basic tooling
This is where most beginners get stuck.

2️⃣ Core Python fundamentals
Start with:
→ Variables, functions, control flow→ Strings (slicing, formatting, methods)→ Lists (loops, sorting, manipulation)→ Dictionaries (key-value thinking)→ Files (reading, writing)
These are the building blocks of everything.

3️⃣ Intermediate skills that matter
→ Sorting and custom logic→ Regular expressions (pattern matching)→ Working with files and systems→ Error handling (exceptions)→ Basic networking (APIs, URLs)
This is where Python becomes useful.

4️⃣ Build real projects
The roadmap is not complete without this.
Examples:
→ Parse real-world data (like logs or HTML)→ Build small automation tools→ Work with files and APIs→ Create end-to-end scripts
That’s how it sticks.

𝗧𝗵𝗲 𝗺𝗶𝘀𝘁𝗮𝗸𝗲
Most people treat learning like watching.
But coding is not passive.
If you don’t practice, you forget.

𝗧𝗵𝗲 𝗼𝗻𝗹𝘆 𝗿𝘂𝗹𝗲
Practice while you learn.
Build something every week.Even if it’s small.Even if it’s messy.

Watch the full video roadmap here
https://youtu.be/8e3Yu_S28s4

3 weeks ago | [YT] | 233

Jean Lee

Most engineers stay stuck in tutorials.These 6 books teach you to build real AI systems.
Tutorials teach you to follow.These books teach you to think.

✅ FOUNDATIONS

Artificial Intelligence: A Modern Approach
↳ Understand the full landscape of AI
↳ Covers search, reasoning, probabilistic methods
↳ The foundation behind everything you are building
amzn.to/4u2uS4c

Deep Learning
↳ The math behind modern AI
↳ Neural networks, optimization, representation learning
↳ What most “AI engineers” skip
amzn.to/4t2o2Lp

✅ SYSTEMS & INFRASTRUCTURE

Designing Machine Learning Systems
↳ How real ML systems are designed in production
↳ Data, deployment, monitoring, iteration
↳ The difference between demos and real systems
amzn.to/41W9jqc

✅ BUILDING MODELS

Build a Large Language Model (From Scratch)
↳ Go from blank page to working transformer
↳ Training loops, tokenization, architecture
↳ No abstractions. You see everything.
amzn.to/4e6kjIT


Hands-On Large Language Models
↳ Visual intuition for how LLMs actually work
↳ Attention, embeddings, vector spaces
↳ Makes complex concepts finally click
amzn.to/4sTbw0r

✅ PRODUCTION AI

AI Engineering: Building Applications with Foundation Models
↳ How to ship AI systems that actually work
↳ Evaluation, scaling, production workflows
↳ Where most AI projects fail
amzn.to/4cVFZp0

Most people learn AI through tools.Very few understand what is happening underneath.
That is the difference between using AI and building it.

Don’t just prompt.Understand the systems behind the prompt.

For a complete roadmap, watch
👉 https://youtu.be/CthJtKKthVA

3 weeks ago (edited) | [YT] | 207