Ramsri Golla Chose a Full-Time Job While Running His AI Products

Running a Full-Time Job and Personal Products Together

Ramsri Goutham Golla kept running his own AI services while holding a full-time job as a lead data scientist at Knownwell. In a post he wrote for his first anniversary at the company, he described that time as the best year of his life, both professionally and personally. For a founder still running Questgen and Supermeme, a workplace with good colleagues was a compelling option in its own right. His career shows a path that moves between founding and employment while keeping ownership of personal products.

Years of Technical Learning and Two Returns to India

The starting point was years of technical learning. After studying electrical engineering and signal processing, he learned deep learning through online courses while working in the United States, then widened his studies to computer vision and web development. When he returned to India in 2018, he planned to try a new business on the back of a relatively low cost of living. It was a move that weighed the skills he already had together with his cost structure.

His early ventures did not go smoothly. He pushed an education technology business in Singapore and explored AI question generation, but the team broke up over conflict between the co-founders. After returning to India in 2021 he stopped actively pursuing a startup, but he kept experimenting with question generation and publishing code.

After that he combined freelance work, courses, and a small software business. In a June 2022 interview he reported monthly recurring revenue of $500 for Questgen and $100 for Supermeme, and he was also earning from Udemy courses. Given the size of product revenue at the time, no single piece of software had yet reached the point of covering his whole living. The arrangement kept several income sources alive while giving the products time to grow.

He also adjusted the scope of the problems he took on. When he built a web service that generated images from text early on, he felt firsthand how hard it was to solve a research-level problem and validate market demand at the same time. Afterward he looked at work that already existed inside an industry and searched for opportunities he could implement with the resources he had. That standard was also an effective constraint when choosing products to run alongside a day job.

The channels for selling accumulated knowledge connected to one another. Questgen's public repository holds instructions for running the question generation code directly, alongside links to the paid course and the web service. Someone who wants to learn how it works can choose the course, a developer who wants to install it can use the code, and someone who wants to skip the setup can use the web service. One research topic led to learning material, open source software, and a commercial product.

The process of building the product was also broken into small steps. The first screen was two text boxes, one for the input passage and one for the generated questions; he built it in Bubble and then added login, saving, and payments in turn. Instead of learning every technology up front to build a finished service, he learned what each feature required and widened the range he could sell.

Products Built by Dividing Roles Instead of Working Alone

He did not insist on building every product alone either. On Supermeme he worked with meme creator Sanjeev NC and full-stack developer Nico Botha, taking charge of AI development himself. He combined the skills each product needed while varying how much he was directly responsible for. His portfolio of personal AI projects came to include both products he ran alone and products he co-founded.

The Supermeme team also kept its collaboration light. Its official description says most communication happens asynchronously around a WhatsApp group. That approach reduces the need for members to be together at the same time. It was a good fit for working role by role while each person kept up their other work and life.

The lighter operating load shows up in Questgen's numbers. In an August 2024 interview he said monthly recurring revenue was $4,000 and that customer inquiries and small fixes took less than five hours a month. That figure refers to Questgen's maintenance and administration at the time and does not include the initial development or the time he spent across his other businesses.

A short maintenance window was preceded by a long preparation. Questgen's official team description says three years of independent research in question generation went into the product. If you take a state of being run on a few hours a month as a given from the start, you miss the learning and development needed to reach it. In Ramsri's case, the period of accumulating skill and the period of maintaining a finished product need to be seen separately.

Customer acquisition followed the same standard of what he could afford. He has said that because he found it hard to spend on marketing in hopes of future growth, he focused on the social media and search traffic he ran himself. Linking product links from existing articles and code was a choice to turn technical activity he had already done into sales.

Testing New Products While Holding a Job

Holding a job alongside a business like this made the relationship with the company matter. In his first-anniversary post at Knownwell, Ramsri thanked Mohan Rao and David DeWolf for trusting him and hiring him full-time even though he was running Questgen and Supermeme. Both also responded in comments with positive assessments of his contributions and of working with him. The public record shows trust between a company that knew about his personal businesses and an employee whose contributions to his day job were recognized.

The employment arrangement and his way of living could be chosen separately. In a post after joining Knownwell, he said a distributed remote team worked well and described bringing his family along to a team meetup. He kept a degree of flexibility in where he lived and in family life while working at the company. The satisfaction he described from the job included the work itself, his colleagues, and the experience of being with family.

While running existing products he also explored new AI applications. AiArtist is a service that turns written descriptions into motion graphics videos, aimed at uses such as social media posts, ads, and short promotional videos. It centers on a flow where a customer enters a prompt and receives a finished video, aiming to reduce the burden of working with a complex editing timeline. The interest that ran through question generation and meme creation now extended to short video production.

This expansion also had a specific technical problem. In a February 2026 post, Ramsri compared building motion graphics programmatically with using a generative video model, and pointed to the difficulty of rendering text accurately inside a video. For AiArtist he described an approach that first locks in the text and composition at the image stage, then generates the motion between them as video. It was development that looked at what fails in the process of producing an actual result rather than at the arrival of a new model itself.

His standard for allocating time to new ideas also leaned toward starting small. He has said to check the response with a light demo product and short social media content, then focus more on the ideas that work. Just lengthening the product list cannot tell you which business deserves more of your time. Small experiments can give a developer who holds a job the grounds for where to assign the next block of development time.

In an interview on July 30, 2026, he reported combined monthly recurring revenue of $6,000 to $7,000 across the three products. He recalled that combined annual recurring revenue had peaked at $100,000 in 2025, and said the newly launched AiArtist had more than 20 paying users. These figures are product revenue he published himself and must be separated from personal income after operating costs and his co-founders' shares. The striking feature of this period is that he was working as a data scientist while holding products at different stages of growth.

Platform Risk and the Standard for Living

Running several products still left a shared risk. Ramsri has said that as large AI services began offering similar features directly, traffic and subscription growth stalled. That experience suggests owning several AI products is not enough on its own to diversify the risk that comes with technological change. Even when the products serve different customers, an expansion of the underlying platform can affect several of them at once.

What he emphasized as a response was the customer's actual use of the product. For Questgen he put forward export to formats that fit various learning management systems as a point of difference. Beyond generating questions, the product covers the stage where customers use the output in their existing systems. That kind of practical difference is also a key criterion when deciding where to spend limited development time.

Time with family was part of the goal as well. In a post about founding a company, Ramsri wrote that you should factor in periods that are hard to get back once they pass, such as the years when you can still travel with your parents or when young children are growing up. This came closer to an argument that you should look at the stage of life you are in when pushing a business forward than to a list of his own family circumstances. He cautioned against the choice of continually postponing other important things until a revenue target is met.

Ramsri's career shows how hard it is to judge a business by whether the founder quit a job. He sold his expertise as courses and products, shared responsibility with co-founders, and chose a satisfying job while keeping his personal businesses. The options his business gave him included both going full-time and running a job and a business together. Which combination to keep could vary with the growth stage of the products, the terms of collaboration, and the life he wanted to lead.