OpenAI Engineer Helps Companies Boost Sales

By age 11, his curiosity expanded from nuts and bolts to software program. He discovered programming languages resembling Basic and Logo and designed easy packages together with one which helped a neighborhood restaurant automate on-line ordering and billing.

Gupta, an IEEE senior member, brings his mixture of curiosity, hands-on problem-solving, and a want to make issues work higher to his function as member of the data science workers at OpenAI in San Francisco. He works with the go-to-market (GTM) crew to assist companies undertake ChatGPT and different merchandise. He builds data-driven fashions and programs that help the gross sales and advertising divisions.

Gupta says he tries to make sure his work has an impression. When making selections about his profession, he says, he thinks about what AI options he can unlock to enhance individuals’s lives.

“If I had been to sum up my general objective in a single sentence,” he says, “it’s that I need AI’s advantages to achieve as many individuals as potential.”

Pursuing engineering via a enterprise lens

Gupta’s early curiosity in tinkering and programming led him to decide on physics, chemistry, and math as his higher-level topics at Chinmaya International Residential School, in Tamil Nadu, India. As a part of the highschool’s International Baccalaureate chapter, college students choose three topics through which to specialize.

“I used to be all in favour of engineering, together with the theoretical a part of it,” Gupta says, “However I used to be at all times extra within the functions: the best way to promote that know-how or the way it ties to the actual world.”

After graduating in 2012, he moved abroad to attend the Hong Kong University of Science and Technology. The college supplied a dual bachelor’s program that allowed him to earn one diploma in industrial engineering and one other in enterprise administration in simply 4 years.

In his spare time, Gupta constructed a smartphone app that permit college students add their class schedules and discover classmates to eat lunch with. The app didn’t take off, he says, however he loved creating it. He additionally launched Pulp Advertisements, a enterprise that printed ads for pupil teams on tissues and paper napkins, which had been distributed within the college’s cafeterias. He made some cash, he says, however shuttered the enterprise after a few yr.

After graduating from the college in 2016, he determined to work in Hong Kong’s monetary hub and joined Goldman Sachs as an analyst within the financial institution’s operations division.

From finance to course of optimization at scale

After two events agree on securities transactions, the financial institution’s operations division ensures that the commerce particulars are recorded accurately, the securities and funds are able to switch, and the transaction settles precisely and on time.

As an analyst, Gupta’s process was to search out bottlenecks within the financial institution’s workflows and repair them. He recognized a possibility to automate commerce reconciliation: when analysts would manually examine knowledge throughout spreadsheets and programs to ensure a transaction’s particulars had been constant. The method helped guarantee monetary transactions had been recorded precisely and settled accurately.

Gupta constructed inside automation instruments that pulled commerce knowledge from completely different programs, ran validation checks, and generated reviews highlighting any discrepancies.

“As an alternative of analysts manually checking giant datasets, the instruments robotically flagged solely the instances that required investigation,” he says. “This helped the crew spend much less time on repetitive verification duties and extra time resolving advanced points. It was additionally my first actual publicity to how software program and knowledge programs may dramatically enhance operational workflows.”

“Whether or not it’s serving to an individual enhance a trait like that or driving efficiencies at a enterprise, AI simply has a lot potential to assist. I’m excited to be a bit a part of that.”

The expertise made him understand he wished to work extra deeply in know-how and data-driven programs, he says. He determined to return to high school in 2018 to check knowledge science and AI, when the fields had been simply starting to surge into broader consciousness.

He found that Columbia supplied a devoted grasp’s diploma program in knowledge science with a deal with AI. After being accepted in 2019, he moved to New York City.

All through this system, he gravitated to the utilized facet of machine learning, taking programs in utilized deep learning and neural networks.

One in every of his main tutorial highlights, he says, was a challenge he did in 2019 with the Brown Institute, a joint analysis lab between Columbia and Stanford centered on utilizing know-how to enhance journalism. The crew labored with The Philadelphia Inquirer to assist the newsroom workers higher perceive their protection from a geographic and social standpoint. The challenge highlighted “information deserts”—underserved communities for which the newspaper was not offering a lot protection—so the publication may redirect its reporting assets.

To determine these areas, Gupta and his team built tools that extracted locations such as avenue names and neighborhoods from information articles and mapped them to visualise the place many of the protection was concentrated. The Inquirer carried out the instrument in a number of methods together with a brand new web page that aggregated stories about COVID-19 by county.

“Journalism was an attention-grabbing drawback set for me, as a result of I actually prefer to learn the information every single day,” Gupta says. “It was a possibility to work with an actual newsroom on an issue that felt actually impactful for each the enterprise and the area people.”

The GenAI inflection level

After incomes his grasp’s diploma in 2020, Gupta moved to San Francisco to hitch Asana, the corporate that developed the work administration platform by the identical identify. He was drawn to the chance to work for a comparatively small firm the place he may have end-to-end possession of initiatives. He joined the group as a product data scientist, specializing in A/B testing for brand new platform options.

Two years later, a brand new alternative emerged: He was requested to guide the launch of Asana Intelligence, an inside machine studying crew constructing AI-powered options into the corporate’s merchandise.

“I felt I didn’t have sufficient expertise to be the founding knowledge scientist,” he says. “However I used to be additionally actually within the area, and spinning up a complete machine studying program was a possibility I couldn’t flip down.”

The Asana Intelligence crew was given six months to construct a number of machine studying–powered options to assist clients work extra effectively. They included automated summaries of challenge updates, insights about potential dangers or delays, and suggestions for subsequent steps.

The crew met that objective and launched a number of different options together with Smart Status, an AI instrument that analyzes a challenge’s duties, deadlines, and exercise, then generates a standing replace.

“Whenever you lastly launch the factor you’ve been engaged on, and also you see the utilization go up, it’s exhilarating,” he says. “You are feeling like that’s what you had been constructing towards: customers really seeing and benefiting from what you made.”

Gupta and his crew additionally translated that first wave of labor into reusable frameworks and documentation to make it simpler to create machine studying options at Asana. He and his colleagues filed a number of U.S. patents.

On the time he took on that function, OpenAI launched ChatGPT. The mainstreaming of generative AI and large language models shifted a lot of his work at Asana from mannequin growth to assessing LLMs.

OpenAI captured the eye of individuals world wide, together with Gupta. In September 2025 he left Asana to hitch OpenAI’s knowledge science crew.

The transition has been each energizing and humbling, he says. At OpenAI, he works intently with the advertising crew to assist information strategic selections. His work focuses on creating fashions to know the effectivity of various advertising channels, to measure what’s driving impression, and to assist the corporate higher attain and serve its clients.

“The tempo may be very completely different from my earlier work. Issues transfer shortly,” he says. “The trade is extraordinarily aggressive, and there’s a powerful expectation to ship quick. It’s been an excellent studying expertise.”

Gupta says he plans to remain within the AI area. With know-how evolving so quickly, he says, he sees monumental potential for process automation throughout industries. AI has already remodeled his core software engineering work, he says, and it’s helped him improve areas that aren’t pure strengths.

“I’m not a great author, and AI has been big in serving to me body my phrases higher and present my work more clearly,” he says. “Whether or not it’s serving to an individual enhance a trait like that or driving efficiencies at a enterprise, AI simply has a lot potential to assist. I’m excited to be a bit a part of that.”

Gupta has been an IEEE member since 2024, and he values the group as each a technical useful resource and knowledgeable community.

He often turns to IEEE publications and the IEEE Xplore Digital Library to learn articles that maintain him abreast of the evolution of AI, knowledge science, and the engineering profession.

IEEE’s member directory instruments are one other priceless useful resource that he makes use of typically, he says.

“It’s been an effective way to attach with different engineers in the identical or related fields,” he says. “I really like sharing and listening to about what of us are engaged on. It brings me outdoors of what I’m doing day after day.

“It evokes me, and it’s one thing I actually take pleasure in and cherish.”