Jessie Xia’s sabbatical journey as a coding intern wasn’t just about learning to code — it was about transforming how she approaches leadership. By stepping into the shoes of her teams, she gained firsthand experience of the challenges and nuances of AI-enabled programming, deepening her empathy and understanding as a leader.
But this journey wasn’t one-sided. For Zhang Wei, mentoring a senior leader like Jessie offered its own set of lessons, from balancing critical feedback with trust to rethinking how to guide someone with years of experience in a completely new domain.
In Part One, we explored Jessie’s first steps into AI programming and the mental shift from CIO to intern. Now, in Part Two, we dive into the defining moments of her internship, the technical hurdles she overcame and the leadership insights that emerged from this unique experience.
Q: What was the most challenging technical concept you faced, and how did pairing with a mentor help you wrap your head around it?
Jessie: The hardest part was admitting I didn't actually understand code. I knew the vocabulary — clean code, refactoring, smells — but I’d never applied them. Standing in front of a real codebase and realizing my theory didn't translate to syntax was intimidating.
Pairing with Zhang Wei provided a safety net. He didn't just show me what to do; he showed me how to think. He planted traps in templates to force me to read and debug the code. Because he made it safe to ask questions, I stopped panicking and started learning. By day three, I was driving the implementation of a sandbox story: Query Agent API.
Q: Did you feel under pressure because of her seniority or were you able to treat her like any other intern? How did you navigate that first moment of critical feedback?
Zhang Wei: Honestly, the pressure was worse before she arrived. I kept thinking: Can I teach this clearly? Will she be bored? Will I be too stiff because she’s a leader? But once we started pairing, the pressure disappeared. We had a shared goal. I focused on explaining the path and she focused on learning it. Titles didn’t matter.
The biggest wake-up call came when we discussed scaling this method to other teams. I kept repeating my concerns about risks and Jessie stopped me: “I heard your concern the first time. If you keep repeating it, it sounds like you don’t trust me.”
It hit me hard. I realized I wasn’t being constructive. I was projecting my own anxiety onto her. Instead of offering solutions, I was just circling the problem. Later, Jessie introduced me to a conflict matrix to analyze the situation. It was a game-changer. I realized that while I thought I was collaborating, my approach was more focused on pointing out issues than finding ways forward. True collaboration isn’t just about raising concerns. It’s about creating a safe space for progress. That moment completely changed how I communicate and work with others.
I realized that while I thought I was collaborating, my approach was more focused on pointing out issues than finding ways forward. True collaboration isn’t just about raising concerns. It’s about creating a safe space for progress.
I realized that while I thought I was collaborating, my approach was more focused on pointing out issues than finding ways forward. True collaboration isn’t just about raising concerns. It’s about creating a safe space for progress.
Q: If AI is the engine and human judgment is the steering, what is the most important skill someone working alongside AI should have?
Jessie: Analytical clarity. AI is a powerful engine, but it’s blind to the why.
In the SPDD workflow, the developer's value has shifted from writing to reviewing and directing. You need to:
Define intent: Use tools like Aupro (our context library) to build project memory and structure context.
Review the blueprint: It’s easier to catch a logic error in a structured prompt than a bug in 1,000 lines of code.
Own the responsibility: AI can outsource repetition, but it can’t outsource responsibility.
For me, this is empowering. The soft skills of analysis and alignment are now the hard skills of the AI era.
Q: For anyone looking to switch lanes and learn a new skill like AI-assisted programming, what is the very first step they should take to set themselves up for success?
Zhang Wei: The first step isn't coding. It's aligning your intent. Try the "One sentence challenge." Can you explain exactly who you are solving the problem for, and what the value is, in one sentence? Also, define the boundaries — tell the AI what not to do.. AI is a lever. It should amplify your clarity, not your impatience. If your intent is clear, AI multiplies your ability. If it’s vague, it just amplifies the mes
Q: What advice would you give to someone who feels too far along in their career to start from zero for the chance to be a beginner again?
Jessie: I’d share my week two mantra: "Sharpening the axe pays off."
You aren't starting from zero. You have system intuition and years of pattern recognition that help you spot when a solution doesn't feel right.
Check your ego: There is massive freedom in being a humble intern. Once you admit you’re there to learn, the pressure to be an expert vanishes.
Do it for the empathy: You can’t lead a team through a habit shift from a 30,000-foot view. You need to feel that day one headache to understand how to help your team through it.
Embrace the risk: The biggest risk isn't being a beginner; it's staying an expert in a world that has already moved on.
Q What was your biggest takeaway from this mentorship experience?
Zhang Wei: My biggest lesson wasn’t about tools or techniques. It was about driving change.
I used to think sound logic and correct methods would convince people. But change triggers psychological defenses, not rational acceptance.
Jessie taught me that people need safety and control to embrace change. Instead of pushing advanced tech, I need to make users feel confident that they can do this and it’s okay to make mistakes. Real change doesn’t come from persuasion. It comes from collaboration, and collaboration starts with empathy.
You can’t lead a team through a habit shift from a 30,000-foot view. You need to feel that day one headache to understand how to help your team through it. The biggest risk isn't being a beginner; it's staying an expert in a world that has already moved on.
You can’t lead a team through a habit shift from a 30,000-foot view. You need to feel that day one headache to understand how to help your team through it. The biggest risk isn't being a beginner; it's staying an expert in a world that has already moved on.
Q: What was your biggest takeaway from this internship experience?
Jessie: This experience taught me so much, but three lessons stand out:
Intent is the new syntax: Years of experience as a business analyst (BA) and project manager are my greatest assets. If you can think clearly, AI can help you build.
Productivity must be governable: We aren't just looking for fast code, we’re looking for AI-first software delivery (AIFSD); a repeatable, scalable engineering system.
Connection multiplies: Individual effort is an accelerator, but team connection, pairing with mentors and aligning with BAs and tech leads is a multiplier.
Jessie’s journey from CIO to coding intern is a testament to curiosity, humility and the value of lifelong learning. Immersing herself in her team’s work allowed her to gain not only technical skills but also a deeper appreciation for the challenges they face and how best to support them. For Zhang Wei, the experience offered a chance to reflect on mentorship and the importance of fostering collaboration and trust.
This story is a reminder that some of the most valuable insights come from stepping outside your comfort zone and seeing the world through a different lens. It’s through these experiences that we challenge what we think we know, reimagine what’s possible and realize that growth isn’t just about moving forward. It’s about seeing the path differently altogether.
Disclaimer: The statements and opinions expressed in this article are those of the author(s) and do not necessarily reflect the positions of Thoughtworks.