Ronnel Maclean Pinto

Moneyview didn’t hesitate to use Nano Banana for illustrations in its IPO prospectus. Here are some interesting tech stats from FY25 ⤵️ • Revenue: ₹2339 cr • Tech infra (A): ₹65 cr • Tech salaries (B): ₹117 cr • Tech salaries / Tech infra: 1.8x • Tech spend (A+B) as % of revenue: 8% PS: Numbers are rounded for better recall. #moneyview #ipo #tech #analysis

4

1 Comment

SEA companies still living 2021? (not judging with a few conversations) Spoke to a CTO recently (Singapore market) and he mentioned that AI is not in their agenda. They see it as just another feature upgrade on the product (Yikes!). They are looking to hire manager to only focus on managing the team, and not be involved in coding (world is changing faster, probably haven't used cursor so far). At this pace, I won’t be surprised if they’re left behind very quickly, especially when many companies are already scrambling to catch up in the AI race. Then we have US and China going all out on AI. History speaks for itself. 🥲 #SEATech #AI #Leadership

6

1 Comment

Last week marked my farewell to AWS after 3.5 years. I got challenged to write this post without any AI so here goes nothing! I joined as a bright-eyed fresh grad, this was my first "big girl" corporate tech role. Back then, cloud felt very abstract, having never seen the inside of a data center and interacting with apps that "just works". While I knew some system design theory from my studies, I couldn't link it to real world applications. AWS changed that. I got exposed to customer innovations, tackling cost optimisation, performance improvements, Well-Architected patterns and more. The opportunity for constant global travel also brought cultural immersion and real-world growth that I never expected to experience growing up. I thought it would be nice to wrap up my time here with 3 key takeaways from AWS that helped me tremendously: 1. Customer Obsession: Making sure to start with the problem, then working backwards from it. Doesn't just apply to tech, but also with everyday life. 2. Embrace change: Tech is constantly evolving at AWS, there are new service features every week, where sometime even customers know it first. 3. Prioritising genuine relationships: My biggest takeaway? The peers and mentors who became true friends and inspire me every day to be better in all aspects of my job. Thank you to all of my colleagues and leaders who have made an impact and shaped not just my career, but me! Like all good things in life, this initial chapter has come to an end, and I am excited for my next adventure. Quoting Lesson 42 from a book titled “Paul’s Life Hacks L500+”, “𝘚𝘰 𝘭𝘰𝘯𝘨 & 𝘵𝘩𝘢𝘯𝘬𝘴 𝘧𝘰𝘳 𝘢𝘭𝘭 𝘵𝘩𝘦 𝘧𝘪𝘴𝘩!”

463

56 Comments

A candidate interviewing for a Senior Engineer @ Meta was asked to design a rate limiter. Another candidate at Google's L5 loop got hit with the same question. I've been asked this three times across different companies. Rate-limiting questions look simple until you add one layer of complexity: – Add distributed rate limiting? Now you're dealing with race conditions and clock skew. – Add multiple rate limit tiers? Welcome to priority queues and quota management. – Add per-user, per-IP, and per-API-key limits? Your Redis bill just exploded. Here's my personal checklist of 15 things you must get right when building rate limiters: 1. Always do rate limiting on the server, not the client → Client-side limits are useless. They’re easily bypassed, so always enforce limits on your backend. 2. Choose the right placement → For most web APIs, place the rate limiter at the API gateway or load balancer (the “edge”) for global protection and minimal added latency. 3. Identify users correctly → Use a combination of user ID, API key, and IP address. Apply stricter limits for anonymous/IP-only clients, higher for authenticated or premium users. 4. Support multiple rule types → Allow per-user, per-IP, and per-endpoint limits. Make rules configurable, not hardcoded. 5. Pick an algorithm that fits your needs → Know the pros/cons: –  Fixed Window: Easy, but suffers from burst issues. –  Sliding Log: Accurate, but memory-heavy. –  Sliding Window Counter: Good balance, small memory footprint. – Token Bucket: Handles bursts and steady rates, an industry standard for distributed systems. 6. Store rate limit state in a fast, shared store → Use an in-memory cache like Redis or Memcached. Every gateway instance must read and write to this store, so limits are enforced globally. 7. Make every check atomic → Use atomic operations (e.g., Redis Lua scripts or MULTI/EXEC) to avoid race conditions and double-accepting requests. 8. Shard your cache for scale → Don’t rely on a single Redis instance. Use Redis Cluster or consistent hashing to scale horizontally and handle millions of users/requests. 9. Build in replication and failover → Each cache node should have replicas. If a primary fails, replicas take over. This keeps the system available and fault-tolerant. 10. Decide your “failure mode” → Fail-open (let all requests through if the cache is down) = risk of backend overload. Fail-closed (block all requests) = user-facing downtime. For critical APIs, prefer fail-closed to protect backend. 11. Return proper status codes and headers → Use HTTP 429 for “Too Many Requests.” Include headers like: – X-RateLimit-Limit,  – X-RateLimit-Remaining,  – X-RateLimit-Reset, Retry-After This helps clients know when to back off. 12. Use connection pooling for cache access → Avoid reconnecting to Redis on every check.  Pool connections to minimize latency. Continued in Comments...

2,279

69 Comments

Coming back from a long Lunar New Year break usually means facing a mountain of catch-up work. It is also a good time to look at how much of that work is actually engineering versus just writing syntax. In a small team, time is our most expensive resource. If my team is spending most of their time on repetitive boilerplate or manual testing, we are losing. My approach to mentoring has shifted to reflect this. I have stopped focusing on the "how" of writing code. Instead, I am teaching the team to use coding agents to handle the routine heavy lifting. We have reached a point where we can complete complex tasks for a fraction of the usual cost. Sometimes the overhead is less than the price of a coffee. I see this as the next logical abstraction layer for the industry. We are moving away from manual code generation in the same way we moved away from managing physical servers. The role of the engineer is shifting toward high-level system design. In our workflow, I provide the architectural direction and the guardrails, then the team uses the agent to build within those parameters at high velocity. We delegate the boilerplate, unit tests, and documentation so we can spend our mental energy on service interactions and data flow. The outcome is a lean team that operates like a much larger organization. Mentoring in this environment moves the focus from "writing code" to "solving problems." When an agent handles the syntax, an engineer's value is found in the "what" and the "why." #FutureOfWork #AIOps #TechLeadership

4

RTO is one of those problems that feels very solvable. Change the 3PLs, add confirmations, escalate harder. And for a while, it improves. Then it quietly comes back! In most cases, not because the courier failed again but because the conditions that created RTO never changed: - Product info on PDP and actual product had a variation - Inadequate information from a customer's lens - COD confirmation mistaken for delivery intent - Addresses assumed “good enough” None of these decisions feel risky in isolation. Together, they almost guarantee friction. What’s interesting is where this finally shows up. Not at checkout. Not in marketing metrics. It shows up days later inside operations. That’s usually when teams realise this wasn’t a delivery problem to begin with. #Operations #D2C #Marketplaces

20