Gaurav Luhariwala has spent a young career walking into rooms where decisions have become chores. First came online shoppers and brands trying to find each other. Then product teams, KYC systems and Indian banks wrestling with credit rules. Now there are job seekers, awake too late, opening too many tabs. The scenery keeps changing. The irritation is remarkably stable: why must a person do the same sorting work again and again when software could carry the pile?
His present answer is Elip AI, a job agent that lives in WhatsApp. A user tells it what sort of work they want through a résumé, a LinkedIn profile or even a voice note. The system says it reads millions of job posts, filters the ghosts and near-misses, and sends a short, ranked list by morning. Some days that list may contain three jobs. Some days it may contain none. Elip would rather be quiet than fill a screen for the theatrical pleasure of looking busy.
That restraint is the interesting wager. The web trained people to confuse inventory with usefulness: more listings, more filters, more little red counters. Luhariwala is betting that a trustworthy agent behaves more like an editor. It does not simply retrieve. It declines.
“Some days that's three. Some days zero. We'd rather send nothing than nonsense.”
A founder with an editorial past
Before the agent, before Y Combinator, there was a curious combination of code and copy. In 2020, Luhariwala worked as a developer at Cakegift.in, helping move a Laravel site toward a MERN stack and translating design wireframes into working pages. During roughly the same period, he was a founding member and editorial board member of The Crescendo, a business magazine tied to a campus entrepreneurship community.
His published subjects were not light filler. They included the Yes Bank crisis, India’s economic slowdown, foreign-investment rules, the agrarian crisis, migrant labor during lockdown, legislation and maternity benefits. The magazine twice named him Young Writer of the Month. He also spent two years as president of an entrepreneurship development cell associated with MNIT Jaipur, later returning to speak to a new group of students about his founder journey.
The pattern is worth noticing without turning it into prophecy. Writing makes a person choose what belongs in the frame. Product work makes the same demand, then charges interest if the choice is wrong. Luhariwala moved between both crafts early. He learned to ship an interface and to arrange an argument.
His next roles pulled him closer to the unruly middle of startups. At Caarya, he worked in a founder’s-office product position. At Tropod, the remit stretched across product and growth. In 2022 he founded Shopflix, describing a trail that began with tools for brands and influencers and moved toward a broader community product. The résumé from this period has the charming disorder of someone learning by touching every lever.
One result escaped the campus laboratory. His Y Combinator biography says he built and sold a direct-to-consumer conversion product to more than 30 brands while still in college. The number matters less than the transaction. A student project became something businesses would adopt. The classroom supplied theory; customers supplied objections in complete sentences.
The operating path · 2020 to now
The six-month edit
At Tartan, Luhariwala moved through product and business leadership roles, working on KYC and data products for Indian financial institutions. His YC launch biography credits that work with growing revenue eightfold in a year. It also placed him close enough to banks to see a peculiar bottleneck: even a small change to a credit rule could take months.
A bank’s lending policy is not a tidy paragraph waiting for a red pen. It is a dense system of rules, data, regulatory constraints and consequences. Tejas AI, the company Luhariwala took into Y Combinator’s Winter 2025 batch, proposed a faster workbench. Bank teams could bring together loan records and policies, test scenarios in ordinary language, examine projected effects, and translate an approved policy into executable rules.
The launch pitch was ambitious but concrete: move risk-policy changes from months toward minutes, while keeping the reasoning visible to the people responsible for the outcome. Luhariwala connected the operational delay to a moral aim as well as a commercial one. In his public announcement, he wrote that access to credit should be fair and data-driven. Faster rule-making, in this telling, was a way to help banks distinguish deserving borrowers from bad risk without leaning on stale guesswork.
The company had technical co-founding strength in Bhavesh Tolia, an engineer with Microsoft experience and an IIT Delhi education. The public launch divided the story cleanly: Luhariwala brought product, business and bank context; Tolia brought engineering depth. Together they aimed at one of enterprise software’s least glamorous and most consequential jobs - changing rules without losing control of them.
A different customer, the same impatience
By 2026, the company’s public face had changed. The Y Combinator directory listed Elip AI in recruiting, and its website greeted visitors with “a job agent in your WhatsApp.” The bank work remained visible as company history, but the live proposition belonged to job seekers. It was not a cosmetic rename. The customer, workflow and business model had all moved.
Pivots invite two lazy stories. One declares fearless genius. The other calls every change a retreat. Neither is useful. What can be observed here is simpler: Luhariwala carried a recognizable product instinct across the border. Tejas attacked a decision system slowed by too many rules and too much manual interpretation. Elip attacks a search system slowed by too many listings and too much manual inspection.
Both products try to compress a sprawling field into a smaller, defensible choice. Both sell time, though to different people. Both must earn trust because a wrong answer has consequences. A mistaken bank rule can misprice risk. A bad job match can waste an application and, more quietly, a little hope.
The through-line is not finance or recruiting. It is the belief that people deserve a shorter distance between evidence and action.
Elip’s interface choice sharpens the thesis. It does not ask a job seeker to install another app destined for the folder of abandoned good intentions. It enters a place where conversation already happens. The user can text, send a voice note or call. The agent remembers preferences and says it tunes future searches accordingly. Every match arrives with reasons the candidate may qualify, an honest catch and an option to generate a tailored, ATS-ready résumé from real experience.
The economics are equally direct on the company’s site. Job seekers pay nothing; employers are meant to pay when a hire happens. That aligns the bill with the party that has a recruiting budget, but it also creates a serious test. The system has to protect the candidate’s trust while becoming useful to the company writing the check. Marketplaces have discovered many elegant ways to make one side feel like inventory. Elip explicitly says the candidate is the talent it finds, not the product it sells.
What builders can borrow
There are three ideas in Luhariwala’s path worth taking home. The first is to study waiting. Founders are trained to look for pain, but waiting is pain with a timestamp. Six months to alter a bank rule is legible. An evening lost to job boards is legible. Once the delay is visible, a product can be judged against a clock rather than a cloud of aspiration.
The second is to put automation inside an existing behavior. WhatsApp is not novel, which is precisely its advantage. Novel intelligence behind a familiar door asks less of the user. The interface disappears into an action they already know: send a message.
The third is to value omission. Software usually celebrates output. Agents will need a better ceremony. If Elip sends nothing because nothing fits, the absence must feel like evidence of work rather than evidence of failure. That requires unusually clear receipts: what was searched, what was rejected, why the threshold held. Trust may come from showing the waste that never reached the user.
Luhariwala’s public record also suggests a broader temperament. A former colleague described him as someone who added life to a team, brought a positive energy and pushed peers to go farther. His own career shows a willingness to move between editor, developer, campus organizer, founder’s-office generalist, product leader and CEO. Those are not proof of personality, but they do reveal an appetite for work that crosses the tidy borders of a job description.
Elip is still an early company. Its promise will be tested in unromantic details: whether the matches are genuinely better, whether ghost listings can be identified reliably, whether résumé tailoring stays faithful, whether employers pay, and whether a chat agent can remember nuance without becoming intrusive. The pivot does not answer those questions. It earns the right to ask them in a market where the team believes it has found leverage.
For Luhariwala, the career so far is less a straight ascent than a sequence of increasingly consequential filters. Which product deserves to exist? Which rule should change? Which borrower fits? Which job is real? Which opportunity is worth waking someone for? His companies keep arriving at the same editorial desk, carrying larger stacks of paper.
The best editors know when the answer is three. They also know when the answer is zero. Luhariwala is building a company around that distinction, one quiet morning at a time.