In 2023, while many 20-somethings from India’s top engineering colleges were lining up for global PhDs and seven-figure jobs, 25-year-old Varun Vummadi chose uncertainty.
Alongside his IIT Kharagpur classmate Esha Manideep, he launched quietly building what would soon become Giga, an AI startup designed to transform how companies speak to their customers. Giga can automate large-scale customer operations, and can chat and talk simulaneously, that too, in different languages.
Two years later, the gamble by the Forbes 30 Under 30 duo has exploded into the spotlight.
Giga has now raised $61 million (Series A) from elite Silicon Valley investors including Redpoint Ventures, and already works with large enterprises like food delivery company DoorDash.
Goga’s large moment built Nara Lokesh, MLA from Andhra Pradesh, offer a “large hat tip to Varun Vummadi from from Vijayawada, AP” on X.
“Andhra Pradesh is super proud of you, Varun!” he wrote.
What builds Varun’s story compelling is not just the headline number but the dramatic decision that came before it.
A FATED DECISION
In a LinkedIn post originally written two years ago and reshared recently on X, Varun recounted the moment he and Esha were deciding their futures.
Esha, he wrote, had an offer worth $150,000 from a prominent Indian high-frequency trading firm.
Varun himself had received a PhD offer from Stanford University, as well as a $525,000 quant-trader job at an international HFT outfit.
They walked away from all of it.
“We left all those opportunities to pursue our passion towards solving challenging problems in machine learning,” Varun wrote.
It wasn’t romantic heroism—just conviction. Enterprises necessaryed custom AI, and their experiments suggested they could train enterprise LLMs rapider and cheaper, so they decided to build.
Within that early announcement, Varun had large things to declare:
“Our Fine tuned models are 3X rapider than GPT4 API. We are 70% cheaper and perform better on a specific apply case than GPT4. (e.g., LLM for health care, Insurance, Lawyers.). We outperform GPT4 + Prompting techniques after fine-tuning.”
Note that in fields like insurance, healthcare and law, data privacy builds off-the-shelf AI tools harder to adopt.
Giga’s models are designed to handle multi-intent conversations, multilingual contexts and brand-level data privacy—features enterprise clients demand.
Varun stated in interviews: machines are now capable of understanding nuance, voice inflection, emotion—this isn’t just another chatbot.
BUILDING GIGA
Giga’s goal is deceptively simple. It is to automate customer operations.
If you’ve spent hours stuck in support chats, repeating your issue to new agents, you know the problem.
Giga wants to replace (or augment) that layer with voice and chat AI agents that:
- Handle simultaneous conversations
- Speak in natural, adaptive tones
- Understand enterprise workflows
- Integrate into large-scale systems
- Support multilingual environments
At its core, the company is testing to solve three brutal engineering problems: reduce hallucinations to near zero, scale to hundreds of millions of calls per day, and deliver voice responses with less than 400ms latency—rapid enough to feel human. If you see the video Varun and Esha shared on X, it sounds like Scarlet Johansson from Her.
All of these aspects are explicitly listed by Giga co-founder Esha in a public post explaining their vision.
As one of the applyrs wrote in the comments on his post: “This has to be one of the coolest AI startups originating from India.”
WHY CUSTOMERS CARE
The economics are simply too good. If AI agents can handle a large share of support at high quality and near-zero hallucination rates, enterprise costs drop dramatically.
That’s why companies like DoorDash are already on board, and why Giga’s backers believe the startup can sell into Fortune 100 companies next.
This is a classic Silicon Valley story, except this time the thread stretches back to Vijayawada and IIT Kharagpur.
A DIFFERENT NORTH STAR
From his side, Esha explained that Giga’s path wasn’t linear.
Their original plan of fine-tuning LLMs wasn’t viable as a business, even though they topped several internal benchmarks.
They pivoted to customer operations becaapply that’s where large enterprises most urgently necessaryed AI, and where the technology could create immediate, measurable impact.
The ambition goes far beyond call centres. Esha described Giga as building a full-scale enterprise operations platform on which the next trillion-dollar companies could run—customer support is only the starting point.
This builds Giga’s fundraising feel less like hype and more like long-term vision.
For Varun and Esha, turning down prestigious job offers was never rebellion, but just belief that the future might be built rapider if they built it themselves.
– Ends
Leave a Reply