President @Hebbia | prev @foundforbiz @uber @postmates @IXLLearning | 🤝 @ashliegudsharma
San Francisco, CA
Joined November 2010
- Tweets1.2K
- Following906
- Followers620
- Likes3.3K
Aabhas Sharma retweeted
@sequoia is leading our Series C at a $1B valuation.
We started Blacksmith in 2024 to make CI faster and less painful. In January we ran 5 million CI jobs a week. Today it's over 90 million and growing, and we’re still early.
ALT Line chart of weekly CI jobs run on Blacksmith, growing from 5.4M in January to 90.5M in October 2026.
We partnered with @AfterQuery and @nvidia to teach an open-weight model how to read credit agreements and amendments.
We got within 3 percentage points of leading frontier models on credit tasks at ~1/7 the inference cost.
With domain expertise - knowing what to teach, what a correct answer requires, and which mistakes matter in practice, I expect open-weight models to quickly match or exceed frontier performance on narrow, well-defined tasks.
We’re excited to prove this out in the real world.
@AfterQuery just moved an open model past the cost-accuracy frontier in finance.
We post-trained @NVIDIAAI Nemotron 3.5 Super VL on expert credit tasks with @hebbia.
Perfect solved: 47% to 74%. That beats Sol 6, and lands within 3 points of Astra 6, at about 1/7 the cost.
Link to the full case study in replies.
AI is letting (even) me rip code faster than ever before, but it doesn’t mean I have immaculate taste, but fortunate to work with @arjmahesh who clearly does (and @FastCompany thinks so too)
woke up to see hebbia design featured in @FastCompany!!
thought i'd share some process for designing max. when we set out to design our harness, we had a ton of ambiguity:
- is a character appropriate for finance?
- should it be anthropomorphic ?
- should it be cute in case it fucks up?
- did it make sense for our brand?
- is it going to clutter the interface?
not to mention our early prototypes literally crashed browsers. the metallic inspo we took from TARS and watch design was heavyyy to process.
so why chase it? it would be easier and faster to just... ship without it. our intuition said giving the experience the extra touch would pay off, and where many eng teams might've just ripped it out and moved on if it was crashing browsers—ours doubled down.
early testing showed that not only was the product a massive improvement, but max gave customers a mental and visual anchor to speak to. they perceived max as trustworthy, and also loved fidgeting with it on our home page (lmao).
5 months later & here’s where we are:
- retention is at an all time high
- brand sentiment is at an all time high
- customers refer to max by name
- we have a new core element to build our product and brand upon
now there’s a new set of questions:
- do people in finance want an army or just one max?
- will people still love it in a year?
- what is the shelf life of maximalism?
takeaways thus far:
1. one of the most exciting things about being a designer today is that ai flattens disciplines. product, brand, animation, & marketing can all coexist and compound user experiences in ways that were too expensive to explore even a year ago.
2. this is not revolutionary in any way, but getting obsessed with our industry, its history, and our customers let us craft something that is uniquely ours in a time where cloning is rampant and the tech industry’s conflation of efficiency and minimalism has diminished opinionated design.
we’ll share a full brand case next week.
cheers.
We do it for the customers, with the people we love (and the revenue follows) 🫶🏽
It’s not just about having a record quarter, it’s also who you do it with. Proud of team @hebbia
Three major launches in the last couple months. You could say losing our heads was inevitable?
👻
Aabhas Sharma retweeted
Grok is the only model sitting on the most important dataset in finance...
The global town square.
"@spacexai might be the best-positioned finance lab on earth."
I'm personally very bullish on AI slowly replacing the work of an analyst. Slightly less bullish on everyone with Claude / ChatGPT simultaneously outperforming each other. G gets at this really well in his latest essay.
Aabhas Sharma retweeted
Astra set a new high in our evals: it produced the best decks we've tested and followed the brief 17% more faithfully than the next-best model, while sourcing its claims to the right document 19% more often. For analysts working through dense financial materials, that means decks and answers you can hand to a client and defend line by line.
Aabhas Sharma retweeted
Weeks after launching Max, we’re growing faster than at any point in our history.
In the last quarter:
• DAUs have 3Xed…
• Total LLM usage has 5Xed, across both frontier, and our own post trained models…
• 40%+ of monthly active users use Hebbia on an average day, and 70%+ use it in a given week, with many of our customers still onboarding to the latest product.
The business has never been stronger:
• We beat 7 competitors to win the largest RFP in investment banking, and closed 10 banks shortly thereafter expanding our TAM well beyond our core investor business.
• We hired 100 people since January, doubling our headcount across London, SF, and NYC.
• We launched a new SKU that closed several of the biggest contracts in company history within *weeks* of first customer conversations.
• The majority of our Series B raise remains untouched. We remain remarkably capital efficient.
There’s a lot of noise in the market. This is a very dynamic market that changes rapidly, and we’ve never been more bullish on our market adoption, product leadership and ability to execute moving forward.
And of course, we’re hiring. careers.hebbia.ai
Come build the future of finance.
Aabhas Sharma retweeted
@hebbia is seeing the token explosion that first came to software development, then legal, and will soon inflect every professional vertical. Maybe we’ll soon be using ‘financial engineering’ to describe something outside of LBOs!
Weeks after launching Max, we’re growing faster than at any point in our history.
In the last quarter:
• DAUs have 3Xed…
• Total LLM usage has 5Xed, across both frontier, and our own post trained models…
• 40%+ of monthly active users use Hebbia on an average day, and 70%+ use it in a given week, with many of our customers still onboarding to the latest product.
The business has never been stronger:
• We beat 7 competitors to win the largest RFP in investment banking, and closed 10 banks shortly thereafter expanding our TAM well beyond our core investor business.
• We hired 100 people since January, doubling our headcount across London, SF, and NYC.
• We launched a new SKU that closed several of the biggest contracts in company history within *weeks* of first customer conversations.
• The majority of our Series B raise remains untouched. We remain remarkably capital efficient.
There’s a lot of noise in the market. This is a very dynamic market that changes rapidly, and we’ve never been more bullish on our market adoption, product leadership and ability to execute moving forward.
And of course, we’re hiring. careers.hebbia.ai
Come build the future of finance.
Aabhas Sharma retweeted
One of the most common questions for app layer companies: what happens as the models become more capable & lower cost?
With recent model improvements, Hebbia's DAU grew 3x quarter over quarter.
Weeks after launching Max, we’re growing faster than at any point in our history.
In the last quarter:
• DAUs have 3Xed…
• Total LLM usage has 5Xed, across both frontier, and our own post trained models…
• 40%+ of monthly active users use Hebbia on an average day, and 70%+ use it in a given week, with many of our customers still onboarding to the latest product.
The business has never been stronger:
• We beat 7 competitors to win the largest RFP in investment banking, and closed 10 banks shortly thereafter expanding our TAM well beyond our core investor business.
• We hired 100 people since January, doubling our headcount across London, SF, and NYC.
• We launched a new SKU that closed several of the biggest contracts in company history within *weeks* of first customer conversations.
• The majority of our Series B raise remains untouched. We remain remarkably capital efficient.
There’s a lot of noise in the market. This is a very dynamic market that changes rapidly, and we’ve never been more bullish on our market adoption, product leadership and ability to execute moving forward.
And of course, we’re hiring. careers.hebbia.ai
Come build the future of finance.
Aabhas Sharma retweeted
Today we’re announcing r-1, our new document parsing model. It’s more accurate than our most powerful agentic OCR models, faster, and up to 6x cheaper.
At @reductoai, we spent two years building specialized models for complex visual layouts, tables spanning multiple pages, and key formatting like strikethroughs. We then used everything we learned to build r-1, the first in a new generation of models designed to handle the hardest documents without multiplying cost.
This early preview delivers a 20% lower error rate than our most accurate legacy agentic models and will keep improving with new checkpoints over the next few weeks.
Accuracy is only the beginning. r-1 preview is available at 1¢ per page all-in, with additional volume discounts as you scale. In the near term we’re also going to release r-1 mini, and an auto mode that intelligently selects the right approach for each page.
If you’re currently using another parser, we’re offering up to $5,000 in credits to evaluate and migrate. You can claim the migration offer and learn more about r-1 using the links in the comments.
Happy parsing!
When we first launched Matrix in 2023 we pioneered the grid interface for AI tools, and changed what it meant to work across thousands of documents.
Matrix 2.0 brings together a firm’s structured and unstructured data, reasons across it, repeatedly takes action, and generates the finished deliverable.
We built it to multiply the judgment and execution of your best employees.
Meet Matrix 2.0.
Aabhas Sharma retweeted
We raised a $45M Series B. The money goes to two things: compute, and people. Compute is the easy part.
Everything @useblacksmith is today was built by 10 engineers. Militant about simplification, allergic to sprawl.
The next chapter is harder. Agent-written code still has to become code you can ship, and the infrastructure for that mostly doesn't exist. We're building it.
Hiring in NYC + SF. Link below.
Years of building alongside the world’s leading financial institutions led us here.
Today, we’re introducing Max: the product Hebbia has been building toward from day one.
The biggest launch in our history.
TRIGGERED
cc @apapirovski @anirbankundu
