These cofounders left Amazon to launch their own business. Big Tech behaviors don’t always translate, and we discovered this the hard way.
ahamedabubakkar35@gmail.com 
These cofounders left Amazon to launch their own business. CREDIT: Andy Ratsirason and Shalini Aggarwal, Tenafli.
Discover How These cofounders left Amazon to launch their own business. 50-year-old San Jose, California CEO Shalini Aggarwal and 37-year-old CTO Andy Ratsirason served as the basis for this as-told-to article. After leaving Amazon at different periods, the two rejoined as cofounders of the AI business Tenfali.
- In order to create an AI firm, two former Amazon employees discuss what they had to unlearn.
- Working with minimal resources and a frugal approach was the hardest hurdle.
- They were able to cut expenses, expedite research, and lessen the need to increase headcount thanks to AI tools.
They discuss the key procedures they had to unlearn in order to expand their startup
Shalini Aggarwal: After I moved to the US from India in 2015, Andy and I started working together. I worked on the side of product and program execution, whereas he was a development engineer.
Andy Ratsirason: In 2014, I started working for Amazon because I wanted to be a part of the Silicon Valley community. I made an effort to adjust my work to fulfill my lifelong dream of being an entrepreneur. I quit Amazon, returned in 2020, and then departed once more in 2023.
I made some changes in starting a startup after I departed for the last time. I came to the conclusion that I required an additional team member. I contacted Shalini after she began to frequently appear on my LinkedIn feed, liking and commenting on pieces about startups and taking chances. We got back together to cofound Tenafli after she had recently left Amazon.
Aggarwal: When we worked in an enterprise organization, we soon discovered that we were taking several tools and processes for granted. The startup world is entirely different; we have to start from zero and have to unlearn many aspects of our attitude.
The AI explosion made it simpler to leave Amazon
Aggarwal: I worked with Amazon until 2024, and during the latter nine months, I worked on AI initiatives, particularly those related to music suggestions.
Ratsirason Nearly three years into my return to Amazon in 2023, I had to decide whether to take a risk and create my own business or stick with those Big Tech salaries. After submitting my resignation in February, I started to consider my options.
Aggarwal: I made the decision to quit Amazon more gradually. My mother passed away and my father retired during the COVID pandemic. I could see how lonely my father was, and finding something to do with his days was the hardest obstacle. The seed was sown by that.
As I worked on personalized AI recommendations, I came up with an idea for a device that could operate as an AI companion for senior citizens, organizing activities according to their regular daily patterns and making unique recommendations.
With the AI boom and my 50th birthday approaching, I knew that I would never make the leap if I didn’t. I gave my two-week notice and quit Amazon in September 2024.
The Big Tech tendency of developing before discovering demand had to be unlearned
Ratsirason: When I worked at Amazon, I didn’t really consider the purpose of the product or whether or not it would be useful. Knowing that the buyer was already there, the “build-first” mentality meant concentrating only on creating a high-quality product.
After months of development, it became evident that this kind of thinking wouldn’t work for an early-stage startup when, on launch day, very few people showed up naturally for a few weeks. That served as a reminder that, in the absence of demand, shipping is not the end point.
After that, before overinvesting in product, we started talking to customers earlier and conducting minor distribution tests, such as waitlists, collaborations, and community postings.
We had to retrain how to be economical after leaving Big Tech
Aggarwal: We applied to and were approved into a number of startup resource programs, such as Google’s Cloud Credits program, Nvidia Inception, and AWS Activate.
Ratsirason: We were given a few thousand AWS credits, but we lost nearly all of them in the first two months before we discovered it. We overprovisioned capacity and left some resources running longer than planned during AI testing, which quietly added up to expenditures over time.
After identifying the problem, we simplified some parts of the architecture to fit our stage, added cost monitoring, set up AWS budget warnings, and included shutdowns in our testing checklist.
We also purchased a tiny local machine to conduct some of our AI tests in order to further cut costs. As a result, we only use the cloud when it is absolutely necessary for scale, managed services, or production.
We were able to spend more time with customers by using AI for research
Aggarwal: We can devote our time to visiting summits, forums, and programs where we can learn about the work of other professionals in the industry because to the time we save with AI. We also spend extra time conducting interviews with potential clients.
Ratsirason: In order to stay up to date on the most recent developments in the industry, we used to spend hours reading lengthy publications and conducting research. We were drawn away from shipping and client conversations by the heavy and slow cycle.
AI is now used to scan a vast amount of content, highlight the ideas that are most pertinent to us, and summarize the few articles that are worth reading. We’ve learned to spend only where learning occurs since we have to be frugal. Develop the simplest thing that can be useful after consulting with users.
Because of AI, some roles are no longer necessary
Ratsirason: AI performs a lot of the coding for us using the specifications we set out, acting as a junior engineer.
Aggarwal: We’ve also learned where scaling won’t be necessary. I am aware that hiring a user interface designer is not necessary. With AI’s assistance, I can swiftly draft something if I understand its criteria and get comments on it.
Ratsirason A few years ago, it would have taken a lot more money and personnel to launch a functional version of our product, Agefully. All we need are subscribers and two engineers. Being a part of this AI era makes me happy.
Although they were helpful, our Big Tech backgrounds sometimes hindered us
Aggarwal: We’ve observed how things operate on a large scale and what procedures we can put in place to prevent future chaos. Overcoming the belief that infrastructure and tools are easily accessible has been the largest drawback. We weren’t properly allocating our energies until we figured out how to deal with that.
Ratsirason: The name itself has some good weight because it comes from Amazon, but it can also work against us. People may conclude that because we worked for a large corporation, we lack the resources necessary to manage a tiny business.
Overcoming the fear of shipping something flawed was the most difficult part. We came from Big Tech, so we were used to high standards and believed that anything less would drive away customers. We discovered that creating something people don’t need is the true danger in early-stage startups, not having sharp edges.
About The Author
