Co-Founder & CEO · Farsight · New York
Samir Dutta
He spent his nights formatting pitch decks at Evercore. Then he built the machine that does it in seconds - and got finance to pay for it.
He spent his nights formatting pitch decks at Evercore. Then he built the machine that does it in seconds - and got finance to pay for it.
Picture a junior banker at 3 a.m., nudging a logo two pixels left on slide 41 because the managing director said it "felt off." Samir Dutta was that banker. The difference is that he kept the grudge, went back to first principles, and turned it into a company that roughly thirty finance firms now pay to make the 3 a.m. version disappear.
Dutta is the co-founder and CEO of Farsight, a New York outfit that does something most "AI for finance" pitches only gesture at. It does not open a chat window and wish you luck. It builds the actual deliverable - the pitch deck, the DCF model, the buyer list, the confidential information memorandum - inside Excel and PowerPoint, formatted to a specific firm's house style. The output looks like the firm made it, because as far as the client is concerned, the firm did.
That insistence on the finished artifact, rather than a clever conversation about the artifact, is the whole bet. Generic chatbots produce text you then have to wrestle into a template. Farsight produces the template, filled, branded, and compliant. The work that used to eat a weekend comes back in seconds.
Dutta did not arrive at this from the outside. He read computer science at MIT, then went where the ambitious analytical types go: investment banking. As an analyst at Evercore he lived inside the deck-and-model machine. Then, as a growth equity associate at General Atlantic, he saw the same manual drudgery from the buy side, where the spreadsheets are different but the late nights are identical.
Two vantage points, one conclusion: the most expensive, best-credentialed people in the building were spending their hours on formatting and reformatting. The judgment that firms actually pay for was getting crowded out by the mechanics of producing the thing the judgment lives in. He had also, before all this, co-founded a music-sharing startup called WeHear as its CTO - so the founder reflex predated the finance detour.
The detail that gives it away: Farsight tailors output to each firm's brand and style guide. Anyone who has survived banking knows the real tyranny is not the math - it is the font, the footnote, the exact shade of corporate blue.
Dutta did not build it alone. Farsight came out of MIT in 2022 with three co-founders, each covering a different flank. Dutta brings the finance domain knowledge - he has actually filed the memo at 2 a.m. Noah Faro, the CTO, brings machine learning from Amazon and the MIT Media Lab. Kunal Tangri, who runs operations, brings ML pedigree from Hugging Face and MIT CSAIL. Finance fluency married to genuine ML chops is rarer than the LinkedIn population would suggest, and it is the reason the product talks like a banker instead of a demo.
In June 2025, Farsight announced a $16 million Series A led by SignalFire, with RRE Ventures, Link Ventures and K5 Ventures joining. The more telling line on the cap table is the angel list: strategic investors from Blackstone, Oaktree, Searchlight and Bank of America. When the people who run the workflows you are trying to automate write personal checks, it tends to mean the demo survived contact with someone who knows exactly where the bodies are buried.
The traction backs it. In 2024 Farsight grew revenue tenfold and expanded its customer roster fivefold. The clients span the corners of high finance that share the same pain - investment banking, private equity, hedge funds, wealth management. Different mandates, same midnight spreadsheet.
It is worth dwelling on what Farsight refuses to be, because the refusal is the product. The reflex in enterprise AI has been to bolt a chat box onto an existing tool and call it a copilot. You ask, it drafts, you copy, you paste, you fix the formatting, you fix it again. The human stays on the hook for the last and most thankless mile - turning a plausible draft into something a managing director will actually put in front of a client.
Farsight's premise is that the last mile is the point. The platform processes unstructured data and produces the client-ready artifact directly inside the native tools - Excel for the model, PowerPoint for the deck, the PDF viewer for the memo. It is tuned to each firm's brand, priorities and compliance posture rather than offering a one-size-fits-all template. The work is delivered end to end, not piecemeal. A buyer list, a valuation analysis, a full pitch deck: produced whole, in the formats finance houses actually demand, in seconds rather than overnight.
That positioning explains the customer math. Firms do not adopt Farsight to experiment with AI; they adopt it because it removes a cost they have been quietly eating for decades. The pitch is legible to a CFO in one sentence, which is rarer in this category than it should be.
Selling automation into investment banks, hedge funds and private equity carries an unforgiving constraint: the data is among the most sensitive in the economy, and the buyers are paranoid by both temperament and regulation. Farsight leans into it rather than apologizing for it. The platform is built around enterprise-grade security - SOC-2 compliance, secure and on-premise deployment options, role-based access control, encryption, and clear data-retention policies. In a world where a leaked deal memo can move markets and end careers, those are not features tucked into a footer. They are the price of admission, and Farsight treats them that way.
In December 2025, Forbes named Dutta to its 30 Under 30 Class of 2026 in the AI category - the same list that, this year, was unusually thick with Indian-origin AI founders. The citation was simple: an AI platform that automates financial workflows for investment banking, private equity and wealth management. The honor confirms what the cap table already implied, that the people watching this space think Farsight is doing something other people are not.
Dutta's Forbes nod did not happen in isolation. The Class of 2026 was notable for how many Indian-origin founders it placed in the AI category - a cohort of builders turning research-lab fluency into companies that sell into specific, unglamorous industries rather than chasing general-purpose hype. Dutta fits the pattern precisely: an MIT computer scientist who picked the least fashionable corner imaginable - the formatting tax on financial deal-making - and treated it as a serious technical and commercial problem. The unglamorous corners are often where the durable businesses hide, because nobody builds a demo for them.
There is a quiet discipline in that choice. Plenty of AI founders pitch a future. Dutta pitches a Tuesday: the associate who would otherwise rebuild the same model from a prior deal, the analyst staring down a buyer list at midnight. Solve the Tuesday well enough, often enough, across enough firms, and the future arrives on its own.
Strip away the funding headlines and the thesis is almost stubbornly practical. Dutta wants the most critical and tedious workflows in finance done end to end - full slide decks, complete models, finished memos - inside the tools people already open every morning. Not a copilot that drafts a paragraph. A system that hands back the whole deliverable so the humans can go back to the part of the job that requires a human: deciding what the numbers mean.
It is a narrow ambition stated plainly, which is its own kind of confidence. There is no talk of replacing the analyst. There is talk of giving the analyst their weekend back. For an industry that has quietly accepted burnout as a rite of passage, that is a more radical pitch than it sounds - and Samir Dutta, who once lived the rite, is betting the firms are finally ready to retire it.
Where to read more & follow along