Himanshu Verma

Swiggy eyes up to $12.7 billion valuation for mega $1.4 billion-plus IPO; launch likely in early November (caution: 70% is going to bank accounts of existing investing only 30% is going into business development)

1

Swiggy can hit 600-700 inr demand is huge got 4 lots in #IPO and bought some in open market. Zomato ipo was around 50-60 now it’s 250. Both Zomato India and Swiggy are bellwether stocks for #india’s affluent and middle class economy as current 100M affluent with spending power expands to 300-400M and middle class around 300-400M expands to 600-800M their valuation can be 10-20x. As pyramids invert game and playbook will flip. This is a duopoly markets and we love duopoly firms

2

Dhoni! The nightmare to Hotstar and Jio Cinema. At the end of this post, I have two links where the Hotstar and Jio executes tell us why Dhoni makes their life so interesting and so difficult. Don't miss to watch. The problem is traffic. Dhoni can take it to astronomical levels. One common solution to this problem is autoscaling. But... hmm... that just does not do magic. Autoscaling does not work when there are traffic spikes i.e., sudden increase in taffic. For prominent cricket matches, hotstar observes spikes of about 1M users per minute adding to the number of active users. A star batsman gets out and the traffic drops by several millions in that minute. This kind of spike is harmful for backend services. Funny problem for hostar is that if users press the back button (instead of just closing the browser), the traffic goes to home page. If homepage was configured for lesser traffic, it gets into trouble. There could be more challenges. What if AWS has no capacity (in a specific AZ) to provision servers to you? So, autoscaling does not always work. Say, you use step-wise scaling, i.e., say, 10 servers at a time with the objective to scale up to 800 from 100. This is too slow. 1 M requests per second, 10 TB of video bandwidth per second (75% of total internet bandwidth available in India) - These numbers are mind boggling! Even the ISPs can get choked delivering this kind of traffic. Streaming services face these challenges all the time. Royal wedding, IPL finals, Trump's activities, and so on. How do you prepare the platform for such spikes? Hotstar uses 3000 machines with 36 CPUs and 72GB RAM, spread across different regions, just to generate the load that hits the platform for load testing. The testing they do to face the spikes is called "Tsunami" testing. A breaking point is calculated for each service. This number is then used to decide the scale of infrastructure. How to solve this problem, then? 1) Use buffer to account for spikes.  2) Pre-warm the infrastructure. So, the uptime is low.  3) Find the right metrics on which scaling can happen "anticipating" a spike, not after a spike has happened. For example, when drinks break ends, expect high load.  4) Use a variety of EC2 instance types. This diversity (memory intesive, compute optimized, etc) can save you if the change in traffic pattern requires a specific resource type.  5) Decide which of your services are P0 and which are less important. Upfront declaration such as that would allow you to reduce resource usage on less important services (and make them available for P0 ones). For example, movie recommendation service is less important than live video delivery esp., when Dhoni is batting. Finally, I note that hotstar and Jio prepare harder for IPL finals than our cricket teams! Watch talks: https://lnkd.in/gkUx27hS and https://lnkd.in/gME-RkxV

18

1 Comment

Companies fail, founders don’t. <excerpt> Wishberg, a wishlist site launched in 2012 by Pravin Jadhav and Kulin Shah, struggled and was acquired by Freecharge in 2014, which Snapdeal bought in 2015. Jadhav was then brought on board at Paytm to start Paytm Money for stock trading, and eventually started his own, called DhanHQ. Shah worked at Acko before founding Onsurity, an employee health benefits platform. India’s internet ecosystem is full of stories of founders starting up, shutting down, being acquired, of being forced out, before they build something that works. Many founders who exit their businesses reinvest in new ventures. Freecharge’s Kunal Shah has funded over 200 startups so far. Venture capitalists also often back founders who have tried something new and failed. India’s startup ecosystem is still innovating and is in the process of maturing. As they say: companies fail, founders don’t. </excerpt> Source link in the comment.

12

4 Comments

My Role in Swiggy's Journey to IPO With India’s leading on-demand food delivery platform, Swiggy became the country’s fastest unicorn in June 2018. I joined the Serviceability team, the core of Swiggy's customer pre-order journey. Our work ensured that any restaurant shown on the app was actually serviceable, factoring in dynamic elements like delivery partner availability, demand fluctuations, and external factors like weather and festivals. With Swiggy's massive scale, each app open triggered serviceability checks for around 2,000–3,000 restaurants, handling up to 100,000 requests per minute at peak. This made serviceability a compute-heavy system, requiring optimizations to accurately calculate distance, time, and delivery service promises. Key Achievements: 1. Caching: Implemented extensive caching for distance calculations, scaling to 200 million requests per minute. 2. Time Prediction Models: Integrated models for assignment, prep, wait, last-mile, and delivery time. 3. System Scaling: Enabled system to handle 100,000 requests per minute within 80ms. 3. Directional GeoFiltering: Replaced Radial GeoFiltering with a rule-based cluster service to improve serviceability accuracy. 4. Debugging Tool: Built a tool for Ops, reducing on-call debugging time by four hours per day and empowering tech support team Working at Swiggy and contributing to its success story as it reached IPO status has been a proud milestone. It’s inspiring to see our efforts in shaping a reliable, efficient customer experience now part of the company's journey on the public stage. #SwiggyIPO #ProudMoment #TechForGrowth #StartupJourney #FoodTech #EngineeringAtScale #ServiceabilitySolutions #TechMilestones #ScalingTech #SwiggyJourney #InnovationInDelivery

161

3 Comments

Arjun Vijay Imagine ordering a Dosa Pan at 00:45 AM because your old ones handle broke, or a Mixer Grinder at 6:45 AM when your current one suddenly malfunctions. Picture yourself needing a mosquito bat in a remote Airbnb late at night, or the convenience of having 10kg bags of rice delivered straight to your door without carrying the carts to parking lot neither carrying in lifts, not to mention daily grocery runs and last-minute orders of socks or underwear before heading to outstation. These are all my real experiences, showing how quick commerce is set to completely transform traditional e-commerce. The new era has already begun, and it's just a matter of time before everyone catches up.

1

Famous SDE, EM behavioural question, "how did you work with cross-functional teams" is not about working. It's also not about the tools you use to work together. It is actually about handling vested interests. Why? Before that, lets first understand what is a cross-functional team. These are teams made of people from different functional expertise. For example, Product Manager, Developers, QA, UX, all in one team to solve a business problem. 3-4 developers + 1-2 PM + 2 QA + 1-2 UX Although these individuals are part of this team, but their managers are not same (in most of the cases). Meaning they are still reporting to their managers of their specific functional team. This is where challenges start to show. Each of these have their vested interests. Dev :- Use the latest and shiniest tech QA :- Make the entire product release a bug PM :- Keep squeezing in more and more features UX :- Become the most frustrated user of the product Shouldn't this lead to a better product? No, because while the team is cross-functional, they are only focused on their specific functions, instead of the specific problem because of which they all came together. Cherry on top is their managers are not same. So, how do you handle this situation at work and answer it in interviews. [1] If anyone deviates from original business problem due to their "best functional knowledge", bring them back by going back to business problem. [2] Communicate like you are talking to a 5 year old. Most of the problems start because we assume communication already happened and everyone is aware of the decisions. [3] Keep communication in written, you don't want to hear later, "oh, I meant that..." [4] Give respect to get respect. Escalate about those who keep disrespecting others. [5] If people disagree, understand where they are coming from. Try giving reasoning behind your stand with data. [6] After disagreement if the final decision in not in your favor, don't take it personally. [7] And finally, talk confidently, explain logically. Projects don't fail due to technology, they fail due to people. It gets worse, when people have vested interests with different reporting. If organisations wouldn't have people problem, then many of you and I won't be having the job we have today (we wouldn't need so many people at workforce). That's why behavioural interview are not hypothetical. They extract the best and worst of you from your own stories and serve it in a platter to interviewer. Need help with difficult behavioural interview questions? DM me COACH.

47

12 Comments

🚀 Revenue Growth: Swiggy's revenue surged by 36% in FY24, reaching ₹11,247 crore (vs ₹8,265 crore in FY23). 📉 Loss Reduction: Losses were slashed by 44%, down to ₹2,350 crore (vs ₹4,179 crore in FY23), driven by a 8% reduction in expenses. 📈 Competitor Comparison: Zomato: ₹12,114 crore revenue, ₹351 crore profit Swiggy: ₹11,247 crore revenue, ₹2,350 crore loss 🍽️ Core Business: Food delivery generated $3 billion GOV with 14.3M MTUs and an AOV of ₹428. 🛒 Quick Commerce: Instamart lagged with ₹1,100 crore revenue and a ₹460 AOV, trailing Blinkit’s ₹2,301 crore revenue and ₹617 AOV. 📊 Instamart’s Reach: Operating in 27 cities with 17,000 SKUs through 523 active dark stores. 💡 Outlook: Swiggy indicates the peak of investments in Instamart is over, paving the way for further profitability. https://lnkd.in/gVC6mJAU

3

I’m shocked by the big layoffs at Intuit, and it's even worse that they called out over 1,000 people as low performers. I bet most of them never even got a poor rating before. One of my friends is also impacted by this layoff, which is very much disheartening. If you are impacted; We have some openings at InMobi, Please check out these opportunities and let me know if you’re interested. 𝐃𝐢𝐫𝐞𝐜𝐭 𝐑𝐞𝐟𝐞𝐫𝐫𝐚𝐥 𝐋𝐢𝐧𝐤𝐬: 1. SDE-2 Android: https://lnkd.in/gP5wTQ3J 2. SDE-3 Android: https://lnkd.in/g4kUw-d9 3. SDET-2: https://lnkd.in/gEhX5tPK 4. SDE-3 Backend: https://lnkd.in/gaV4Wwvu 𝐄𝐱𝐩𝐥𝐨𝐫𝐞 𝐌𝐨𝐫𝐞 𝐎𝐩𝐞𝐧𝐢𝐧𝐠𝐬: Visit the InMobi career page to find other positions that match your profile. If you see a role that suits you, share the details with me, and I’ll gladly refer you. #layoff #intuit #recession #opportunity #opening #refer

131

15 Comments

🚨 Google Slashes Map Prices in India by 70%: Too Little, Too Late? 🚨 Google has announced a 70% price cut for its map services in India. But let’s be real—this move comes conveniently after Ola announced its own mapping services. 🤔 This price cut comes just one week after Ola CEO Bhavish Aggarwal offered one-year free access to Ola Maps for all developers. The announcement was made by Google at its developer event, Google I/O Connect, in Bengaluru on July 17. Why now, Google? This price drop feels more like a response to competition rather than a genuine effort to support Indian businesses. After years of high prices, why should we trust that this discount is in our best interest and not just a temporary measure to squash Ola Maps? 💡 Indian companies and consumers deserve better than reactive measures. Let’s support innovation and fair competition, not just price cuts driven by competitive threats. What do you think about Google’s sudden price reduction? Share your thoughts below! 💬👇 #TechCompetition #GoogleMaps #OlaMaps #Innovation #FairMarket

3

2 Comments

Intuit layoffs were disheartening! How can a company suddenly comes up with 1K+ employees as "#underperforms" ? Why there were: No 1-1 with the manager, No performance improvement plan (PIP) , No prior notice What they were doing from past few months? This raises the question on #Employment #Laws and workplace regulations. Labour Law & Regulations needs to be reframed and amended for this longhorn treatment! . . . . #LaborLaws #EmploymentRights #WorkplaceRights #Layoffs #JobSecurity #EmployeeAdvocacy #WorkplaceJustice #HRCompliance #EmploymentLaw #WorkerProtection #WorkplaceEquality #JobLoss #EmployeeWellbeing #WorkplaceSupport #HRManagement

13

Let's talk about quick deliveries today like Instacart, Blinkit, instamart, Dunzo, Zepto and others. Pros: - You get your essentials in 10 minutes. (that's what they claim but only 5-10% of the time, it's delivered with-in claimed 10 minutes) Cons: - We end up wasting more time than saving time (we started using these apps to save time but that never happened) - You end up spending more than double/triple amount than actually required on essentials because their marketing team is running have some or other offer/discount everyday - You end-up ordering junks more than essentials because you are on their app on daily basis - Social Isolation: Relying heavily on delivery services can lead to decreased social interactions and potential long-term social isolation. - Slowly, your life becomes an unplanned mess. - Making delivery partners anxious, always - Putting delivery partners & others lives on risk who all are on road - Environmental Impact: The increased use of delivery vehicles contributes to pollution and delivery often generates unnecessary packaging waste, adding to the environmental concerns. What we do(for more than a years now): - We order our monthly groceries using Dmart at starting of the month (takes 30 minutes every month) (grocery prices, delivery time & delivery charges are reasonable as compared to any other platform, mall or local shops) - Fruits and dry vegetables on weekly basis from local market (takes 30-45 minutes every week) - Green vegetables once or twice a week (no extra effort) Planning is essential part of our lives and in my opinion we should at least plan for essentials(groceries, fruits & vegetables). Though it's completely YOUR choice. Above observations are completely on my personal experiences and may/may not resonate with yours but if it resonates, let others know by interacting with the post :) #life #experience #planning #essentials #quickdeliveries

36

9 Comments

Few people join as a SDE-1 and within a span of ~7 years, they promote to a very senior level(L6 equivalent to Staff Role) in a company whereas same achievement is hard to realize when you're switching companies. So, what makes these employees so special? A few points: - They joined these companies as SDE-1 - Implemented projects end to end - very pro-active in all things: reviews/discussions/coding. - highly biased for taking actions (critical skillset for growth) - were lucky enough to some extent in getting good projects - were hard working enough to find optimizations & propose as new projects in the product - filled with lots of motivation & drive: showing true ownership at the deepest level one could possibly imagine Staying in a company helps you build context over the years, which the lateral hires within few months might or might not be able to grasp. This post is just a reminder: Switching would probably make money but staying in a company with a little bit of luck can do wonders(both money & role). PS: I did 2 switches in my career and I can relate with this post.

550

12 Comments

CapEx spends go over 10% of a company’s revenue—largely for expansion, new business, and sustaining current operations. However, finding technically qualified vendors offering quality output, agreeable pricing, and timely delivery remains a laborious and testing process. Before starting Venwiz, I realized why this process needs to be objective but lacks the means to enable that. Subjective choices are prone to biases, ignorance, and relationships interfering with what should be a transparent and rigorous process for high-spend items. Our goal has always been to eliminate these inefficiencies. Not having an objective vendor selection process can: - Increase project costs unexpectedly - Delay project timelines, causing frustration - Compromise project or end product quality - Cause reputational damage But given information asymmetry and time pressure, it isn't easy to stay 100% objective in ranking vendors. By incorporating empirical data, defining parameters based on first principles—like vetting past experience on similar projects—and utilizing technology, we at Venwiz narrow down the most suitable vendors from an otherwise large universe, close to real-time. That's unseen in most of the manufacturing industry, and it feels special each time. What challenges have you faced while selecting vendors? #procurement #vendors #Projectmanagement #Venwiz #CapEx

33

After creating the Google Maps Platform for India, it's time for us to create our own YouTube. Recently, YouTube stopped monetization on my account saying that my videos and not mine while I have reused someone else's content and uploaded them directly. My channel was the first one (at least up to my knowledge) to create content on low-level system design and now they are saying I am putting a reused content. This is so disheartening to hear that. Even on asking, they did not mention which video is reused. If they are so sure, then tell me which video is reused and I will show them how it is not a reused one. But just banning the account without giving details is bad. It's high time that just like Bhavish Aggarwal created our own Maps, we bring in our own YouTube so that this monopoly of Google can go away where they are dong anything they want with the creators without giving even proper details and reasons. Just yesterday, I recorded one video on system design to put on YouTube but now I am in dilemma about whether I should put it or not. #pathetic #bad

389

55 Comments

IPL: Razorpay's Second Innings 🏏 The title "IPL: Razorpay’s second innings" really intrigued me. What does a fintech platform have to do with IPL? 🤔 It turns out, user behavior during an IPL match, like ordering food or gaming, had an indirect impact on Razorpay. In 2019, they struggled with traffic bursts during IPL matches, which lasted over 10 minutes—something their engineering hadn't anticipated. After this challenging IPL season, they strategized to handle future traffic bursts better. Here are the key problems they identified and solved: Problems: 🧵 Thread Exhaustion: PHP Laravel, their backend, lacked proper DB connection pooling, blocking threads and connections when the gateway took longer. 🚦 Throttling: They weren’t throttling requests, making them vulnerable to DDoS attacks. 🏦 Banking Gateway Backpressure: Banking gateways experienced 20x their normal P99s during traffic bursts. 🚥 Smart Routing Requests: There was no way to identify the load on different banking gateway nodes to route requests to less loaded ones. 📊 Alerts & Monitoring: Their alerting mechanism was too slow, giving them no time for remediation. Solutions: 🧵 Thread Exhaustion: They used ProxySQL as a layer between the backend and MySQL, deploying it as a sidecar with each backend pod for fault tolerance. 🚦 Throttling: Implemented simpler fixed window rate limiting instead of the earlier one which was a leaky bucket algorithm implementation backed by a Redis store, also moved this service to an NGINX-based proxy server as a side car with dedicated cache and new rate-limiting configurations (e.g., requests per route per merchant). 🏦 Easing Banking Gateway Backpressure through Smart Routing: Built a machine learning-based system to route transactions smartly, considering earlier payment success/failure events. This improved success rates by 60% during bursts and 4% during normal periods. 📊 Alerts & Monitoring: Using Victoria Metrics and Grafana, they developed a real-time alerting system for business and engineering KPIs. This is how Razorpay overcame their engineering challenges! For more interesting engineering stories, follow me. https://lnkd.in/gbzx8qen #Engineering #Fintech #Scalability #MachineLearning #DevOps 🚀

9