OpenAI engineer Sarang Gupta helps companies find customers and lift sales

Sarang Gupta's career shows how the engineer's role is changing in the age of generative AI. Today a strong technical specialist is needed not only on a research team but also where a company explains a product's value to the market, measures channel effectiveness and helps customers get results faster. At OpenAI, Gupta works at the intersection of data, marketing and sales, building models that help businesses attract buyers and better understand what actually drives growth.

His interest in such problems began long before OpenAI. As a child in India he liked taking apart and fixing household items, and by age 11 he had switched to software. He studied Basic and Logo, wrote simple programs and even built a solution for a local restaurant to automate online orders and invoicing. This practical outlook only strengthened later: at school he chose physics, chemistry and mathematics, but he always saw engineering not as an abstract science but as a tool that has to work in real life. So after graduating in 2012, he enrolled at the Hong Kong University of Science and Technology in a dual program combining industrial engineering and business management.

This mix of technology and commerce shows in his early projects. While studying, Gupta built a mobile app that helped students upload their timetables and find classmates to have lunch with. The project didn't take off, but it gave him product development experience. In parallel, he started a small business, Pulp Ads, printing advertising for student organizations on napkins and paper items for campus cafés.

After university, in 2016, he joined Goldman Sachs, where he worked on operational processes for securities transactions. His job was to find bottlenecks in workflows and remove them. One notable case was automating trade reconciliation: instead of manually checking spreadsheets and different systems, he built internal tools that pulled in data on their own, ran validation and flagged only problem cases. It was an early example of how doing data work properly frees people from routine and lets them focus on complex decisions.

That experience pushed him deeper into data science and AI. In 2019, Gupta started a master's program in data science at Columbia University, where he focused on applied machine learning, deep learning and neural networks. Among his most important projects, he singles out work with the Brown Institute and the newsroom of The Philadelphia Inquirer.

The team built tools that extracted geographic mentions from news texts (streets, neighborhoods, counties) and visualized the publication's coverage on a map. This helped identify so-called news deserts, meaning communities to which the newsroom paid too little attention. For the news industry, it was not just an analytical experiment but a way to allocate reporting resources more deliberately.

The case explains Gupta's style well: he is drawn not to models for their own sake but to situations where they change processes and produce a measurable effect.

After his master's, in 2020, he moved to San Francisco and joined Asana as a product data scientist. At first he was responsible for A/B tests of new features, and then he led the launch of Asana Intelligence, an internal initiative meant to build AI capabilities into the product within six months. The team managed to ship several features, including automatic project summaries, alerts about risks and delays, next-step recommendations, and Smart Status, a tool that drafted status updates on its own from tasks, deadlines and project activity.

What matters is not only the launch itself but also that the team then packaged its work into reusable frameworks and documentation so that AI features could be shipped faster inside the company.

When ChatGPT arrived on the market, the focus of his work shifted from developing proprietary models to evaluating and deploying large language models. In September 2025, he moved to OpenAI.

At OpenAI, Gupta works on the data science team alongside the go-to-market group. He is responsible for models that help determine how effective marketing channels are, what contributes most to results, and how companies can better convey the value of ChatGPT and other products to customers. This is not the most visible part of the AI industry, but it is one of the most practical: without this kind of analytics, even a strong product is harder to scale.

Gupta's story matters because it reflects a shift inside the AI industry itself. Great value is created not only where new models are trained, but also where they are turned into a clear business tool, where returns are measured and where they are brought to mass use. For OpenAI, such specialists become a bridge between technology and the market. For other companies, they are a reminder that AI adoption starts not with a trendy interface but with a clear task, good data and process discipline.