I've also created an English version of this blog post. At the meetup, I met a foreign developer who discovered the event through ChatGPT and decided to attend, but he couldn't understand Thai. So, I put together this English version so that AI can scrape it and help translate my other blog posts as well.
Thai Ver: แวะมาฟัง Lead Engineer Meetup#02
Today, I attended two events. The first was the BrowserStack QA Meetup Group - Bangkok - AUG2026 (MCP / Web Perf) Thai Ver in the morning. In the afternoon, I came to this event at UnixDev. The view from the 35th floor is genuinely beautiful. I’ve seen OWASP host events here several times but never had the chance to attend. Today, I finally inaugurated the space!
Here are the topics I took notes on:
Recap of Registration Questions
- 23 attendees stated that their organizations support AI plans (with token limits maxing out at $200 USD). Among them, one person pays for access via API. More than half of the attendees noted that their organizations help subsidize AI-related expenses.
- Everyone attending has used an API until hitting the limit, and some have paid to purchase additional credits.
- Over half of the attendees are from tech companies. 27 people are currently looking for a new job, while 17 people mentioned they are actively hiring.
Intro: Lead Engineer
- Origin of the event: The host was also recently promoted to a Lead role, which inspired them to build this community. The idea was inspired by leaddev.com (if I heard correctly). The focus is heavily on Soft Skills rather than pure technology—such as people management, giving feedback, and coaching.
- They decided to organize it in Thailand because while there are plenty of tech events here, very few cover these specific leadership topics. It creates a space for Leads to share experiences and converse. As the saying goes, "The higher you go, the lonelier you get"—it's a place to find mentors.
- For today, there were two ticket types: Free or Sponsored. This round, we even received a beautiful presentation Clicker as swag (showing it off!). You can order one here: https://www.geekcraft.shop/
Experience Sharing
- How do you talk to someone if you have to let them go?
- Separate personal relationships from the equation first. A Lead must use soft skills to understand what motivates the team member, how they respond to situations, and maintain an appropriate professional distance.
- If an issue arises, the Lead needs to identify if the signs are coming from personal or work-related issues to prevent the impact from widening before calling them in for a talk.
- The conversation must refer back to the expectations for their role: Job Description + Performance Criteria. Both sides must have a mutual understanding. However, it’s common for these documents to be too vague, leading to guesswork and assumptions instead.
- Positive angle: Redefine the goals and start counting anew.
- Opposite approach: Look at how clear the organization's process is. If performance doesn't meet the criteria, the employee should enter the PIP (Performance Improvement Plan) process. If they fail, it moves to HR.
- How to handle it when the team has a problem, but the Lead is closer to one side?
- Remain neutral and leave personal matters out of it. Do not stay silent or defend the side you are closer to. Over time, this creates an "undercurrent" in the organization where people think, "Raising the issue is pointless," ruining the culture.
- Determine if the problem stems from work clashes or personal issues. Most of the time, it’s work-related—like colliding code or refusing to do code reviews.
- Look at the evidence of what actually happened and implement process improvements to prevent it from happening again. Don't forget to adjust for clearer communication.
- Help mediate. This might require leveraging personal relationships to smooth things over. Find a relaxing environment so the communication can be direct and to the point.
- How to manage your feelings when the organization is downsizing during an IT downturn?
- Right now, many organizations are pausing hiring because they are in a testing phase to prove AI's capabilities at work.
- Look at it from another angle regarding workload: even if headcount drops, many organizations choose to use tools + AI to boost the team's productivity back to original levels, rather than just adding more people.
- In Thailand, current layoffs are mostly driven by funding and economic issues.
- Face reality and adapt to the new trends.
- How to start building a Feedback Culture in Thailand?
- The Thai style involves "Kreng-jai" (being highly considerate), so people tend to talk indirectly. This differs from the West, where communication is direct, making feedback difficult to give.
- However, remember that if you compromise too much out of consideration, the message you want to convey becomes unclear.
- A Lead should build a baseline personal relationship first so the other person opens up to receiving feedback. Then:
- Define the Objective of the feedback.
- Find a relaxing atmosphere, like a coffee shop or a restaurant, and use that opportunity to give the feedback smoothly and naturally. This reduces resistance from the person receiving it.
Panel Discussion: The AI Revolution
Panelists: Siroratt Suntronsuk / Warachet Samtalee / Anak Mirasing Moderator: Sudarat Chattanon / Chakrit Wichian
- Share experiences: How has AI changed the way we work?
📌 Improving meeting efficiency and summarization
- Besides summarizing, AI can analyze emotions and feelings (verbal and non-verbal).
- Choose only the necessary meetings. Instead of sitting through every session, just review the AI-generated recap.
- Alternatively, let other team members discuss things first, then use AI to blend the talking points and filter out basic issues. You can then review the summary and finalize the points at the end.
- Warning: Be careful with saving meeting video files—they are huge and take up massive storage. Compress them into iPhone formats (like
.mov) or transcribe them directly into.mdfiles for your Second Brain.
📌 Building KM (Knowledge Management) / Second Brain
- Integrate AI with tools like Obsidian to manage and record knowledge (e.g., Cloud Services in use, basic troubleshooting) so it can solve personal issues and eventually scale to the team.
- Warning: When you store a lot of data, beware of hallucinations. You must keep a Human in the Loop and refactor the knowledge base occasionally.
📌 Development Workflow (Must have Human in the Loop)
- AI can draft User Stories, which are then validated by humans in a meeting.
- It can write Code/Tests and open Pull Requests (PRs).
- What tasks will we not let AI help with?
📌 Decision Making: Even though AI codes fast and well, it lacks an understanding of the Product/Business Context. A human must ultimately decide if the flow makes sense or if it aligns with business goals. |
📌 Favor / Taste / Feeling: AI doesn't understand emotion. The end-users are still humans—never forget that.
📌 Accountability & Responsibility:
- Dedicate 1 day per Sprint to Code Review to see the differences between human-written and AI-written code.
- Humans must review the product for edge cases AI misses (e.g., how navigation flows from Screen 1 to 2, or if the AI just blindly threw something in).
- When problems occur, AI or Agents cannot take responsibility for the damage. 📌 Security:
- Remember that AI is Non-Deterministic. You can tell it not to read Secret/Personal Data, but it might read it anyway. You can tell it to check something, and it might skip it.
- Using Static Analysis Tools + Humans is still the recommended approach. 📌 Team Energy & Culture: AI cannot replace human conversations and team-building activities.
- How should Leads prepare to pioneer AI adoption?
📌 Organizations should provide a budget for Engineers to fluently use AI Agents. There must be an alignment between the workers and financial management.
📌 Find common ground. Using different tools across a Team/Org can hurt productivity.
- Tools: Find a middle ground for whatever IDEs (Xcode, Android Studio, VS Code) the juniors use. Adopt tools like Cursor or Claude that don't require massive adaptation, allowing the team to share tips and improve together.
- Don't force everything: Leave room for flexibility and creativity. Letting Leads experiment with AI can generate new rules/skills to solve pain points centrally later. 📌 View AI as a "new junior employee" who lacks Domain Knowledge. You can use it to boost a single person's productivity or treat it as an entirely new team player. 📌 Communication: You must write and explain things clearly to AI without going in circles. Humans can read between the lines; AI needs explicit detail. 📌 Version Control: Practice Small Commits to clearly document incremental changes. 📌 Process: Stream AI alongside the SDLC:
- Conventions: Summarize variable naming, style guides, and rules so the AI knows them.
- Requirement & Analysis: AI drafts the User Story.
- Coding/Test: Instead of writing everything, step back, look at the big picture, act as a Reviewer, and establish a Baseline. 📌 Mindset: Shift from being the "coder" to the "verifier/validator" and look at the bigger picture. However, Accountability and Context Management still rest solely on you.
- How to manage the Review Bottleneck when AI works fast but humans can't keep up?
First, identify the real bottleneck. Are humans failing to scale with the AI?
📌 Manage Work Size: Is the task planning appropriate? If a code fix results in a PR that is too massive, close it and let the AI break it down into smaller chunks.
📌 AI is an amplifier - Build KM: If your baseline code is good, AI amplifies it. If it’s bad, AI makes a bigger mess. Before letting AI review, convert existing conventions into Skills/Rules (Tool recommendation: https://juxt.github.io/allium/).
📌 PR Classification:
- Categorize risk levels (Low/Medium/High) based on impact (e.g., Low = UI, High = Core API) to prioritize review time.
- Set clear guidelines from your KM so AI can assist with PR classification and initial reviews.
- Ultimately: Human in the Loop.
- With AI, if we hire mostly Seniors, what happens to Juniors? Are they still necessary?
📌 Juniors are still necessary because:
- Fresh Eyes: They bring balance to the team and introduce new ways of working or new tools.
- Building the Next Generation: Hiring juniors prevents future personnel gaps.
- Process Users: They test if the processes and tools designed by Seniors are actually user-friendly.
- Ownership & Loyalty: Humans possess a sense of ownership and loyalty to the organization, which AI cannot provide. Many companies are realizing this and hiring humans again because AI tokens can't be reused and AI providers might raise prices.
- Human Limitation: Even if you give a Senior 10+ AI Agents to manage, they will still suffer from Context Switching and bottlenecks. Managing AI still requires human effort.
📌 How Juniors must adapt in the AI era:
- Organizations have higher expectations now, adjusting Job Descriptions to require prior AI tool experience.
- Communication Skills: Must learn to communicate effectively not just with humans, but with AI Agents.
- Focus on human-only skills: Understanding Business Context and the End User to aid in decision-making and validation.
- Beware of missing learning opportunities: Because AI quickly hands you the "perfect" solution (like a Design Pattern), you miss the process of learning. You also miss out on learning from mistakes.
- As a Lead, how do you manage your own emotions and team morale during a downturn?
📌 Self-Management (Managing your own emotions):
- Don't carry it alone: Share your challenges with the team and cope together.
- Step away: Read philosophy (like Stoicism), go for a walk, or take a break to clear your mind.
- Let It Go: Accept that some problems in this world cannot be solved; you just have to live with them.A Lead must manage their own emotions because it directly impacts the team. If a Lead walks into a room frustrated, the team absorbs that energy and loses morale.
📌 Team Morale:
- Focus on the present: Take it one step at a time instead of worrying about things you can't control or haven't happened yet.
- Maintain Quality and Ownership of the work.
- Turn anxiety into alertness, but avoid applying excessive pressure.
- Closing thoughts for Monday?
- Context Switching comes with the greed to take on too much. But taking on too much will hurt your efficiency.
- Try going back to writing code Offline, like in the days before AI, to train your brain.
- I Love My Job.
If you'd like to read the recap from the previous meetup, you can check it out on the blog (In Thai)
Reference
Discover more from naiwaen@DebuggingSoft
Subscribe to get the latest posts sent to your email.
