Ronnie Chong

Rule of thumb: under-promise and over-deliver! If your partners agreed to an under-promised (safe) ETA, all good and ethical. If you managed to deliver more earlier, that's great - your partners will have a pleasant surprise. If you delivered on time what was initially scoped, all good as well. Keep in mind that there's always a game played in the background. Your stakeholders have tight deadlines, so if they have slippage, there's a safety buffer. But there are expectations that they'll make it. So, they push others in exactly the same manner, trying to get tighter ETAs to de-risk. This propagates, aggregates, and makes everything top priority, as if it should be delivered yesterday. In reality, most things are negotiable and often not as urgent as presented.

83

8 Comments

GIFs in notifications are now natively supported on Android (Although with a certain caveat, supported since Android 12 - https://lnkd.in/gUDp9_s2, but work on Android 14 and above 🫠) I haven't seen any apps using it. Popular CRM tools do not support it out of the box could be a reason. However, it could be an interesting addition to your marketing toolkit differentiating your app from others. Implementing it is pretty straightforward. During the creation of your notification, download it to your app's files folder and provide it via a content Uri to the notification builder val notificationManager = getSystemService(Context.NOTIFICATION_SERVICE) as NotificationManager if (Build.VERSION.SDK_INT >= Build.VERSION_CODES.S) { val uri = FileProvider.getUriForFile(applicationContext, "com.example.notification", File(applicationContext.filesDir, "ponds-wb-variant.gif")) builder.setStyle(NotificationCompat.BigPictureStyle(). bigPicture(Icon.createWithContentUri(uri))) }; Android adoption is also much better these days with 40% of affluent Tier 1/2 India audience having a phone running Android 14

161

7 Comments

What’s new at tekFinder? Over the last two years, we’ve rethought what tech recruitment can do for good. We’re still deeply committed to building strong tech teams, but we’re now just as serious about broader impact. This is why we joined the Pledge 1% movement. And while I continue to help organisations fill critical roles with skilled, values-aligned candidates. What’s different (and important to understand) is that every hire through tekFinder now plays a part in strengthening digital capacity for grassroots charities right across Australia, via tekFoundation. What's more, larger charities can also work with us on their IT recruitment, knowing their typical hiring spend is actively helping smaller organisations thrive, with no extra cost attached. Why do we do it this way? We're a social enterprise and the first of its kind in Australia: tekFinder is for-profit. tekFoundation is not-for-profit. The two feed one another - and that’s by design. We exist to help values-led organisations find exceptional IT talent - and uniquely, every placement through tekFinder supports our charitable arm, tekFoundation. This means that while you’re building your internal tech team, you’re also funding technology volunteers for smaller charities that simply can’t afford digital help. That could be anything from web development or data dashboards through to basic IT support. If you want to see the impact in action, have a look at tekfoundation's linkedIn page or website. Interested in supporting social impact through your next IT hire? I’m always open to a conversation. #TechForGood #SocialImpact #DigitalInclusion #CharitySector #Recruitment

43

2 Comments

I’ve been building products for over a decade, and this is the most significant shift I’ve seen in how teams work. For most of my career, product building followed a pretty familiar shape: 🟦 PMs define 🟨 Designers craft 🟩 Engineers build You could draw a neat triangle—with handoffs, ownership lines, and clear lanes. But AI has changed all of that. Now: — Engineers are shaping how systems behave, not just how they scale — Designers are prototyping inside the model, not just on the screen — PMs are tuning prompts and debugging UX through model interaction The lines are blurring because the product isn’t static anymore. It’s dynamic, adaptive, and sometimes weird—and shaping it means working across boundaries. This isn’t chaos. It’s convergence. And the teams that embrace that—the ones who collaborate beyond title and territory—are the ones shipping the future. #ai #productdesign #engineering #startups #ux #productmanagement #buildinpublic #leadership

122

12 Comments

I chatted with John Kutay recently about Strands Agents and what our "model-driven approach" to building agents really means, such as swapping out RAG and prompt pipelines for Retrieval as Tool or "RAT" (coined by my colleague Anoop Deoras!). Let the agent's model ask for what it needs to know, instead of proactively stuffing the context window with unnecessary information Strands Agents: https://strandsagents.com/ Podcast links: Striim https://lnkd.in/gWhdzktk YouTube https://lnkd.in/g5GSYCGr Spotify https://lnkd.in/gHS8rdz6

187

4 Comments

Innovation is a big deal at Certn, especially on the Product and Engineering side. We’ve been leaning hard into AI - not just because it’s the cool thing to do, but because it’s actually useful, both for our clients and for our teams. For clients, it helps us scale without losing the quality and consistency they rely on. And for our development teams, it removes so many of the old blockers so they can spend more time solving real problems and building cool stuff. It’s a fun time to be in software - the tools are evolving fast, and we’re using them to ship better, smarter products.

39

1 Comment

Your title doesn’t change the rules of interviews. Even Staff Engineers and Tech Leads get grilled on DS/Algo—especially at top MNCs and product companies. Yes, you’re designing systems. Yes, you’re driving architecture and mentoring teams. But that doesn’t mean you’re done with data structures and algorithms. In the past, you might’ve gone through 10+ interviews before landing the right offer. Each interview itself became a practice session. But now? Fewer interviews. Higher stakes. Less room for error. You won’t get multiple attempts to “warm up.” And if you’re serious about switching, or just staying sharp—you can’t afford to ignore DSA. Reminder: Your current job title won’t protect you from a binary decision in an interview. You either solve the problem or you don’t. Don’t let overconfidence blindside you. Prep smart. Stay sharp. The fundamentals still matter.

35

2 Comments

Don’t bend. Don’t break. Don’t back down. To everyone in my feed saying “software engineering is dead”: relax. The only thing that’s dead is your curiosity. AI isn’t coming for developers who actually like developing. It’s coming for the people who chased the easy money, “thought leaders” who’ve never shipped a pull request, and keyboard warriors whose most complex build was their LinkedIn profile. This industry isn’t a free ride and it never was. Most careers worth having are competitive. Ours just has cooler merch and more caffeine dependency. If you love this craft, now is your time to sharpen your skills, learn the tools (including AI), and show the quitters why you’re still here. Don’t bend. Don’t break. Don’t back down. #DontBendDontBreakDontBackDown #QuittersGonnaQuit #CodeWarriors #EasyMoneyIsOver #AIAndDevelopers #ShipOrHushUp #NoCloutWithoutCode #KeepCodingKeepWinning #NotForTheWeak #LongLiveSoftwareEngineering

1,508

218 Comments

A contrarian’s perspective into Deepseek’s R1 impact on OpenAI’s moat. I’m an engineer, so my perspective is grounded in what truly pushes the boundaries of technology. While Deepseek’s ability to drastically cut costs and speed up inference times is undeniably impressive engineering, I find it less revolutionary than many claim. Optimizing GPU usage and leveraging existing hardware techniques are valuable but essentially incremental gains rather than entirely new breakthroughs. Why? Because the Deepseek team already knew what it was optimizing for. If OpenAI hadn’t pioneered the direction of chain-of-thought reasoning, Deepseek might not have known which path was worth accelerating in the first place. It’s similar to sports, where breaking a world record can seem impossible—until someone does it. Then suddenly, many athletes surpass that once unbeatable mark. The real leap, and where OpenAI remains laser-focused, is in pushing the frontiers of reasoning itself—developing advanced chain-of-thought systems and novel intelligence architectures. That is the truly hard problem, and the one that will always be tough to replicate or outdo. Speed and cost reductions undoubtedly help, but the far more challenging (and moat-defining) achievement is creating new forms of AI-driven reasoning, not just squeezing more performance out of existing hardware. Deepseek’s R1 is a commendable feat of engineering, but in my view, it doesn’t fundamentally alter the landscape of the biggest breakthroughs on the horizon.

82

9 Comments