future startup logo

Engineering Reliable Solutions For Complex Problems Like Grocery, Transportation, and Logistics At Chaldal With Tejas Viswanath, CTO, Chaldal

Flight Expert Credit banner


Chaldal Co-founder and CTO Tejas Viswanath on how his passion for programming has led to what he is doing today, his journey from National University of Singapore to the founding of Chaldal, his work at Chaldal as the CTO and how he approaches his work as a source of fulfillment, how Chaldal is using engineering to solve complex problems like grocery, transportation, and logistics in Dhaka, how Chaldal’s tech team operates, its user-first technology philosophy, how Chaldal is using machine learning and other tools to improve delivery time and overall efficiency of its operation, how Chaldal hires engineer, and the connection between Chaldal’s overall strategy and its ride-hailing business Chalao, culture at Chaldal, major tech trends, biggest opportunities for markets like Bangladesh and why technology should make our lives simpler and better.

Future Startup

What is your background? How did you end up at Chaldal? Please tell us about your journey to what you are doing today.

Tejas Viswanath

I’m an engineer at heart. I started programming on an old Intel 80186 with 512 KB of RAM that was manufactured in 1982. Somebody had abandoned this computer, and I booted it up and learned to program it at the age of 13. That’s when I fell in love with computers, as it seemed so natural.

My first software application was a DOS-based salary calculator spreadsheet written in C, which was deployed in my father’s manufacturing company. Eventually, I ended up automating quite a bit of my father’s company, including building out inventory, purchase order systems and dispatch systems by the age of 17.

I was in India at that time and was given an all-expenses-paid scholarship by the Singapore Government to continue my studies in Singapore for both high school and college.

During my time at the National University of Singapore in 2005, I created my first internet start-up, Nineo.com, to help provide jobs for students, that went on to become the largest student database in the region at that time.

Eventually, I moved on to work for SigFig in San Francisco, an online finance portfolio tracker and “Robo-advisor” company, where I spent 5 years and was running the company’s tech infrastructure (across 4 data centers).

We were doing high-profile projects with Yahoo! Finance, AOL Daily Finance, CNN Money and USA Today, and I was in-charge of architecture, performance, scaling, data centers, hardware and securing all our users' assets, worth more than $2 Billion at the time.

While in SigFig, I met Waseem, the Chaldal CEO, who wanted some help creating an e-commerce website that he was going to use to sell grocery. I helped him out part-time, just spending a few hours in the weekend.

One day I got a call from him saying that the project is growing fast, we’re getting too many orders, and that he needs a lot more tech support.

At that time, Bangladesh was a clean slate, where we saw an opportunity to bring the best of the world’s technology and connecting them up to create something at a whole new level. Along with his friend, Zia, this initiative then became Chaldal.com.

I’ve always had a preference to work with startups as opposed to big companies such as Google because as you’re creating the platform you work on the way you want it to be. You also have a lot more control over the destiny of a startup.

In a big, established company, you’re bound by what is already created; there are a lot of politics that will restrict your growth and thinking. A good metric to ask is “will my presence in this company alter its course significantly?”. If the answer is no, i.e. if you’re dispensable, then you’re wasting your time.

For me, work isn’t really “work”, in the traditional sense of that word, where you come into the office, get your day’s work done, and head out at 6 pm. The way I approach it can be best described through a Japanese concept known as ikigai, variously translated as “a reason for being”, or “a reason to get up in the morning”.

[su_divider top="no" divider_color="#adacab" link_color="#edde29" size="1"][/su_divider]

A Message From Flight Expert

Flight Expert banner

It's all about your journey!

Flight Expert is Bangladesh's first online travel agency (OTA). Our goal is to unlock all the facilities of the online travel agency (OTA) so that you, our customers, can easily book flights or hotels online from anywhere without leaving the comfort of your home or office. We believe when our customers win, we win. Hence, putting your needs first is our business. Every day we work hard, seek new ideas and new ways of doing things to make sure that you experience the best trip of your life. You can learn more about us and avail our service here.

[su_divider top="no" divider_color="#adacab" link_color="#edde29" size="1"][/su_divider]

Future Startup

Please tell us about your work as the CTO of Chaldal. How do you approach your work?

Tejas Viswanath

For me, work isn’t really “work”, in the traditional sense of that word, where you come into the office, get your day’s work done, and head out at 6 pm. The way I approach it can be best described through a Japanese concept known as ikigai, variously translated as “a reason for being”, or “a reason to get up in the morning”.

On a good day, when I wake up, my mind is filled with ideas, I can feel the excitement in my nerves and I can’t wait to get started. The difficult part is in choosing which of those ideas to focus on for the day and pushing everything else out of my mind.

I’m driven by my vision to see a highly connected world. Our world is very inefficient today because systems don’t talk to each other; and when they do, they don’t do it very well, for a lot of non-technological reasons (politics, walled gardens, incompetence, etc.).

I believe we have maybe executed 5% of what we’re capable of, given our current set of available technologies. It is possible to create an environment so remarkably different it will appear magical.

It takes a lot of grunt work to work on each piece of the vision, but every day I know I’m a step closer.

Future Startup

Could please explain Chaldal from a technical point of view? Although you provide grocery delivery service, you are a tech company at the core. How does tech enable different pieces of your operation?

Tejas Viswanath

Chaldal is a tech company to start with, and most people laugh at the idea of technology being needed to sell rice and dal.

Selling groceries at scale is a very complex problem. Currently, we’re processing more than 1500 orders daily, and some days we’re processing more than 2500. This number is doubling every 6 months. Each order contains on average 15 items, and these items can be in any one of our 8 warehouses.

We have about 8000 products in our catalog and some items may need to be purchased from local markets. Our software helps us put all this together and get it into the customer’s hands within the hour.

The margins on selling grocery are very small, and we need to do this very efficiently for us to remain in business. This is hard, physical work and it is very easy to make a mistake. We also need to react to exceptional situations, such as a product being damaged or missing. Doing this with pen and paper is impossible when you’re targeting 1-hour deliveries. Technology enables us to provide these services.

As to “why grocery?”, it is a constant part of all our daily lives; everyone needs it. Building such a fundamental service will give us the opportunity to interact with our customers very often and will let us build more services on top of this platform. Grocery shopping hasn’t changed much over the last few decades because it is the hardest problem to solve in e-commerce.

Given the population density of Dhaka and the proliferation of technology through all sections of society, we believe we’re in a unique position to be able to centralize grocery delivery as a service.

When operating at scale, we will also have an opportunity to do some good. We believe that hunger is a distribution problem; and that it is possible to reduce wastage and cost, to make quality food available to everyone.

Future Startup

How big is your tech team? How does your tech team work at Chaldal?

Tejas Viswanath

Our tech team currently has 8 members. We’ve hired very slowly. Other companies that have attempted to do this elsewhere in the world normally have many dozens of engineers at our stage. Yet we’ve managed to get to our current numbers by being very efficient and by hiring great people.

The team is spread across 3 offices – Dhaka, Singapore and San Francisco. Culture is very important for us; life is too short to spend your time with someone who isn’t a friend.

We have a no-meetings policy; we don’t want to waste people’s times with pointless bureaucracy. Our company is founded on the realization that engineers are the best people to build our products, so they’re given the authority to make decisions; we don’t have non-technical managers. Putting good engineers under a non-technical manager is the surest way to the kill engineering creativity.

We’re also very cognizant of stress. I believe that 6-hour focused workday can take us very far, so we do not expect people to work through the nights. A lot of companies pressure their engineers into working well into midnight, or through weekends, without realizing that the long-term output of these people will reduce.

Sleep and relationships are important; when these get affected you burn out faster, which leads to engineers working with lesser intelligence over the long term.

There may be days when our servers go down or there is a critical issue, that may require all-hands-on-deck, but this is generally the exception rather than the rule.

Selling groceries at scale is a very complex problem. Currently, we’re processing more than 1500 orders daily, and some days we’re processing more than 2500. This number is doubling every 6 months. Each order contains on average 15 items, and these items can be in any one of our 8 warehouses. We have about 8000 products in our catalog and some items may need to be purchased from local markets. Our software helps us put all this together and get it into the customer’s hands within the hour. The margins on selling grocery are very small, and we need to do this very efficiently for us to remain in business.

Future Startup

What is your philosophy when it comes to technology and usage of tech? For many companies, tech is an enabler for their users and for others, it is ‘the thing’ such as Facebook where users have minimal control. When you design a tech solution, how do you think?

Tejas Viswanath

Our core philosophy is simple, “do what is right”. Every situation that presents itself has a “right” response, and we believe that each of us knows internally what this “right” thing is.

Often, we do the “right” thing, yet we sometimes don’t for many reasons. While this may seem like a commentary on moral issues, it applies to tech just as well. It is what is used to decide on how we respond to deadlines or critical outages. Or whether we should use a library for a feature or build our own.

We’ve found that a lot of non-technical managers do not care about this, and simply want their work done at whatever the cost. What might then happen is engineers might be forced to take shortcuts to appease their managers and produce an unoptimized buggy system. This will then need to be addressed at a later point, but now more time is needed for investigation and refactoring. This will require more engineers, more servers, and more funding. All while re-doing the same work, instead of moving forward.

On the other hand, if we simply “do what is right”, we might take a week more to build something, but then it simply chugs along in production. We still must face challenges as the product requirements change, or when the traffic scales up, but at least we don’t have to do patch-work because the original author was either impatient or was pressured into doing bad work. This culture goes all the way up in our company, even when it comes to non-technical matters such as ops and HR. This I believe, is one of the core strengths of our team.

Facebook once made this famous quote to reflect their internal culture – “Move fast and break things”; we believe this has done more harm than good, as young engineers now feel justified in producing sloppy work in the interest of time. Within our company, we use a modified quote, “Move fast and don’t break things”.

As to “why grocery?”, it is a constant part of all our daily lives; everyone needs it. Building such a fundamental service will give us the opportunity to interact with our customers very often and will let us build more services on top of this platform. Grocery shopping hasn’t changed much over the last few decades because it is the hardest problem to solve in e-commerce.

Future Startup

Technology-wise, what is the strategy?

Tejas Viswanath

Our strategy is to move more and more of our code into what are known as functional programming languages, which are closer to mathematics, to reduce the number of bugs in our code, and increase their efficiency and iterate our systems faster.

It is a new way to think, and all our new projects are now being worked in this new paradigm, with our older projects being slowly migrated over.

It is a challenge for new engineers to learn this, but once when do, it is like getting a software upgrade for your brain. They write better code from that point on. Everyone who has done this within our company has had positive feedback on this process. It helps us release projects like GoGo Bangla with just 2 engineers instead of 30.

We have realized that when technology works to the best of its potential, it is virtually indistinguishable from magic. This is another reason we only recruit top talent; people who aren’t scared of doing what it takes. We have a saying within our company, “To a magician, there’s no such thing called magic”.

Future Startup

What are the technological challenges for Chaldal? One challenge that we could see from the outside is related to inventory management - Chaldal is still probably manually track what is available and what is not. What are other challenges?

Tejas Viswanath

We have two main challenges – increasing on-time delivery record and making sure we can deliver all ordered products. Both are difficult problems to solve.

Several issues can arise during deliveries – traffic, delivery man getting into accidents, vehicle failures, some customers take a long time to collect their products etc. Our software needs to get better at being able to predict such delays and plan accordingly.

We also need to be able to plan our contingencies better in the face of unpredictable events. At the same time, we will need to keep the entire system very efficient, so that we can keep the delivery cost low.

Keeping our deliveries efficient cannot be solved perfectly using today’s computers, as it is a combination of 4 Travelling Salesman Problems (TSPs). A TSP is a well-known example that belongs to a class of problems known as NP-complete; solving them efficiently is a fundamentally unsolved problem in computer science. The best we can do is to come up with an approximation using heuristics.

Our inventory tracking is completely automated, however, sometimes products go missing or are found damaged without a replacement. Some vendors may delay their supply of these products. Our handling of such situations need to get better, and we’re working very hard on getting our fulfillment rate close to 100%.

We’re also very cognizant of stress. I believe that 6-hour focused workday can take us very far, so we do not expect people to work through the nights. A lot of companies pressure their engineers into working well into midnight, or through weekends, without realizing that the long-term output of these people will reduce. Sleep and relationships are important; when these get affected you burn out faster, which leads to engineers working with lesser intelligence over the long term.

Future Startup

You just launched Chalao, your ride-hailing service, how does that connect with your overall tech ecosystem within Chaldal? One way to look at it probably is that you will connect both operations at some point on the logistics end to have greater efficiency. But then and again, there are challenges like the service level agreement and other. How do you see these two pieces of your operation interact in the future?

Tejas Viswanath

We’re working on connecting this up. Once this connects, we’ll be able to introduce “ASAP” deliveries for certain areas based on the availability of drivers. Not all products can be delivered by riders on motorcycles (e.g. 10kg bag of rice), so we’ll design the product accordingly.

Another possibility is to be able to deliver any product available in the city to anywhere, with a listing on Chaldal.com, so we can indeed become a 1-stop shop. We will be announcing more updates on this in the coming months.

Future Startup

How do you find and hire tech talents?

Tejas Viswanath

As with any other technology company, or really any company, finding talent is the hardest part. Historically, we’ve hired through direct referrals from our network, with a few through job advertisements.

We’re highly selective on who we hire into the team. A team of 2 A-grade engineers can often accomplish more than 40 B-grade engineers. A-grade engineers are generally those who are detail-oriented, care about what they do and put in the effort in doing things the right way; very predictably they come with a long history of success.

Most tech companies do not realize the overhead of having a large team; once you go beyond 15, you’ll need to hire engineering managers, architects, dev ops personnel etc. You’ll need to have planning meetings, meetings to discuss how planning meetings will work, HR meetings, strategy meetings etc. It’s not uncommon for the output of 1 engineer to come down to 50% of their normal productivity as the company grows and starts to lose cohesion.

Doing this very early in a company’s lifetime will require expensive capital and the need to raise large funding rounds, which will further pressure a company to deliver better results back to their investors to pave way for the next funding round, which will result in everyone in the company always being under pressure and will lead to an inferior product or service, which will eventually lead to bad results. This is a vicious cycle we don’t want to be caught in.

While the productivity drop isn’t completely avoidable as we scale, we’ve attempted to delay this problem for as long as possible.

We hired our first engineer one year after Chaldal.com started, the second engineer 18 months after. 3 years after, we hired 3 more engineers and 3 more after 4 years. This has allowed us to be the most capital efficient vertically integrated e-commerce business in recent times, something we’re well known for in investor circles.

We have realized that when technology works to the best of its potential, it is virtually indistinguishable from magic. This is another reason we only recruit top talent; people who aren’t scared of doing what it takes. We have a saying within our company, “To a magician, there’s no such thing called magic”.

Future Startup

What does it take to excel at Chaldal as an engineer?

Tejas Viswanath

First and foremost, being a nice person. Our observation has been that the nicer someone genuinely is, the greater the intelligence they possess.

And when I say intelligence here, it doesn’t mean knowing which API or function to call with what parameters – anyone who has enough free time can learn that. But it is the ability to come up with smart & creative solutions and being useful to those around them. This is something, not every engineer (even those at high levels) is good at. Besides, why would you want to waste your time hanging out without somebody who isn’t nice?

That is the first metric, an absolute requirement. It’s very easy to tell when somebody is genuinely nice – they’re nice to everyone around them, and they make you feel warm and comfortable. Others put in the effort of being nice to only those who control their destiny in some form – interviewers, managers, CEOs etc. During interviews, this easily shows, as they will always tell you what you want to hear. They can be easily trapped into contradicting themselves.

Another thing we gauge is whether the person wants to work with us for the right reasons; i.e. will they find some level of fulfillment or satisfaction when they work with us.

Sometimes candidates and engineers may not really know what they want, and when such situations arise, instead of trying to retain them, we actively encourage them to find what they truly want, even if it means leaving us.

This is important because when someone is working with you for the right reason, they are automatically driven – nothing more needs to be done to further motivate them. They will find and solve problems that need solving.

Our aim is to create an open environment where people can come in and find satisfaction by contributing towards something. We simply try to direct this force towards the company’s vision.

On the technical side, we test their problem-solving ability; normally this reflects the net experience gathered by that individual thus far. It tests what patterns they’ve encountered till now, and how effectively they’re able to apply them to solve unfamiliar problems, which determines their level within the team. Given that we focus as much on knowledge & learning as the delivery of work, engineers continue to level up as they work with us.

Future Startup

A few exciting things that you are building at Chaldal? And there is a lot of talk about AI and machine learning, are you using any of these technologies at Chaldal? If yes, how does that work?

Tejas Viswanath

The most exciting thing we’re building is “ASAP” deliveries, where we try to deliver the product to you as soon as possible; based on where you’re located, this might be as quickly as 15 minutes. We don’t yet have a release date, but when this service is launched, it will make all our deliveries a lot more efficient.

This is a complex problem as we need to prepare the baskets in our warehouse within 5 minutes, while the delivery logistics team plans the delivery route and gets the delivery vehicle prepared. There are so many moving parts in our system, and everything must work together perfectly to make this happen.

The engineering team is made of hard-core functional programming wizards, and we’ve invented a patented technology to very quickly create operational processes, i.e. a way to convert your thoughts into a fully functioning operational business. We use this to launch new software very quickly.

For example, the entire complex backend for the Chalao ride-sharing service was designed within a week by 1 person. This is something that would normally take months to launch. Ultimately our goal is to enable non-engineers to be able to produce technology-enabled businesses like Chaldal and Chalao.

We’re getting deeper into machine learning by using it to predict stock requirements by-the-hour based on numerous factors, like location, day-of-the-week, holidays, weather, strikes etc. The better we can write this algorithm, the better we can stock products, the cheaper we can acquire this stock for, and more savings we’ll be able to pass on to customers.

We’re deploying machine learning into our operations so that we can predict operational failures (e.g. late delivery or missing items in the basket) long before they happen so that we can take corrective action preemptively. We have enough data at this point to predict very interesting things, for example, that employees make mistakes picking the wrong ketchup bottle just before their lunch time as they’re hungry.

We have two main challenges – increasing on-time delivery record and making sure we can deliver all ordered products. Both are difficult problems to solve. Several issues can arise during deliveries – traffic, delivery man getting into accidents, vehicle failures, some customers take a long time to collect their products etc. Our software needs to get better at being able to predict such delays and plan accordingly.

Future Startup

From a technical viewpoint, what are the future plans for Chaldal; robots delivering grocery, AI, what are the plans?

Tejas Viswanath

We’re planning an expansion to other parts of the country, by extending our logistics network to other cities and even rural areas, as we hope to create a truly connected digital Bangladesh.

We want to enable easier trading of commodities all the way back from farmers and importers and help connect them better to retailers; this will mean both better quality and prices for customers.

We want to be able to create the world’s model e-commerce city, where anything available in the city can be discovered and brought to your doorstep in as little time as possible, for as cheap as possible.

As our orders double, we will need to use robots to prepare baskets and manage our dispatch zone, as manual processing is too slow and error-prone.

Down the road, as AR technologies mature, we’re looking into making grocery shopping more interactive by creating a virtual supermarket in your living room. This will help users browse our selection in a 3-D world as opposed to a 2-D screen. This is already possible with current technology, but VR headsets are clunky right now for consumer use, so this needs to evolve.

Drone deliveries are still a bit far away; we’ve considered this before, and the biggest problem is the battery densities on the drones don’t allow us to deliver large products. Not to mention that it is illegal to operate drones in the country.

First and foremost, being a nice person. Our observation has been that the nicer someone genuinely is, the greater the intelligence they possess.

Future Startup

What do you think about the overall technology industry? There is incessant talk about AI and related technology, what are a few other tech trends that you see going big that will have a greater impact on markets like Bangladesh?

Tejas Viswanath

AI is largely a pop-culture term, and those who really do know what happens behind the scenes normally don’t refer to these technologies as A.I, rather by their more specific names. It is simply a collection of algorithms and statistical methods that have been available to us for decades.

Google has made deep learning very accessible to the average developer via TensorFlow, and other companies have followed suit; this is causing a lot of buzz. Some applications of this “A.I” are quite funny; it appears that we’ve found ourselves a golden hammer, and we’re now looking around for nails to hammer. Which is fine, there will still be a few good ideas that come out of it, but it is important to recognize that “A.I” as it stands today is more buzz than substance.

The blockchain, in my experience, is another such technology, where there is more buzz than substance. The blockchain itself is a marvelous invention; we just still haven’t figured out how to use it. Again, when something good emerges out of this, it will come out of nowhere and will revolutionize our industry. I just think that it still hasn’t happened yet.

In markets like Bangladesh, which are still very much dependent on manual pen-and-paper processes, the biggest opportunities still lie in digitizing the economy. People underestimate how much potential lies here, as this is “boring” work.

Young engineers tend to want to work on the newest and shiniest technologies that are emerging; this attitude itself is great. But it needs to be combined with a measured investigation into what these technologies really bring. Does it make someone’s life better? Is this something somebody needs right now? These are important questions to ask. Ultimately, we believe that technology is an equalizer; something that can help people break through class divides, and we have a long way to go in realizing this vision.

I do hope Augmented Reality tech takes off and becomes more accessible to all. The real impact of this technology will be when it can help people telecommute better. Our cities are overcrowded causing massive traffic jams and real-estate spikes. With good AR, a significant part of the working population can telecommute, while being as equally effective as being in the same office with their coworkers.

Future Startup

How do you personally use technology given the fact that many of today’s technologies such as smartphones or Social media are designed to take advantage of human cognitive weaknesses and addictive in nature?

Tejas Viswanath

I simply think we’re going through a growing phase, and we’ll soon outgrow these problems. Too much importance is given to social media, as opposed to development projects.

In the last few years, it's not that we’ve stopped working on development projects; its just that social media has gotten a LOT more coverage, since everyone is a part of it, and hence find it more interesting. This is more of a perception bias than an accurate picture of what’s really happening.

While smartphones are addictive in nature, it has also made life very comfortable. The addictive aspect simply brings to the surface what was already a problem – that people feel a certain emptiness in their lives. This applies to all of us, universally.

As a society, we have already started talking about this; governments and social networks have become more aware of the impact of social media and the price of always being plugged into electronics. This conversation itself is a good first step.

Personally, I try to not get caught up in this conversation. Our journey ahead in terms of acquiring knowledge and understanding is very long and every bit is immensely exciting. Technology is but a tool that I use to make my life simpler.

We’re planning an expansion to other parts of the country, by extending our logistics network to other cities and even rural areas, as we hope to create a truly connected digital Bangladesh.

-
This interview was done via email.

In-depth business & tech coverage from Dhaka

Stories exclusively available at FS

About FS

Contact Us