
Engineering @TryPallet | Previously @scale_AI
San Francisco, CA
Joined March 2013
- Tweets158
- Following87
- Followers124
- Likes918
Vidhur Kumar retweeted
“Our customers keep asking us to buy more agents every single quarter after they start.”
@SushanthRaman and @kevinzhang on the growth behind @TryPallet's pivot.
Vidhur Kumar retweeted
.@TryPallet's confidence was restored during its hardest stretch. @SushanthRaman and BCV Partner @kevinzhang on what it took to get there, and why the desire to survive can take you far.
youtu.be/sM_L9zC1UYw?si=qEMD…
Vidhur Kumar retweeted
If you played Age of Empires II or WoW as a kid, I'd lowkey trust you with my life. You were running economies and raid groups at f*cking 12 years old.
Vidhur Kumar retweeted
America winning the AI race also depends on how quickly we can diffuse intelligence into the physical economy.
Having a high pain tolerance is not optional when it comes to building generational businesses 😤
Sometimes building an outlier company means walking away from a good one.
The first version of @TryPallet was working. But Sushanth Raman believed it needed to change to deliver truly ambitious impact for the global supply chain. He pushed ahead with a pivot despite resistance from customers, teammates, and investors.
@kevinzhang and @SushanthRaman discuss what drove the decision, what it cost, and what came next.
youtu.be/sM_L9zC1UYw?si=dWh0…
Vidhur Kumar retweeted
Seems like they put Decart before Dehorse
EXCLUSIVE: Anthropic has walked away from a potential $6 billion acquisition of AI startup Decart bloomberg.com/news/articles/…
Vidhur Kumar retweeted
Two million shipments a year at 99% accuracy is a distributed systems problem wearing an AI costume.
Under the hood: typed, functional, event-driven plumbing. Immutable versions. Every shipment is data, every side effect is a liability.
Boring on purpose. Proud of this one.
We're proud to partner with @onelineage, the world's largest temperature-controlled logistics company.
Today, @trypallet AI agents process two million of their shipments a year at 99% accuracy, helping them scale without compromising the high-touch service they're known for.
Vidhur Kumar retweeted
. @1Password the CLI grant window should show which items/fields are being requested, not just which process is asking. Right now there's zero visibility into what a script will access. Per-item transparency would be huge!
We've had so many instances in candidate onsites when we've brought on people who spiked hard in one interview and failed another interview.
100% of these hires have been absolute killers. Look for spikes in your hires.
Vidhur Kumar retweeted
boy if i was chairman of the openai board this would be music to my ears
Rent the intelligence, own the relationship - how Sierra's Context Engine helps you build a competitive advantage in AI sierra.ai/blog/context-engin…
$480M ARR, growing 20% YoY, have a fuckton in the bank.
Still decides to burn the boat and focus entirely on AI.
Really wonder if the “adapt or die” mindset is being taken a bit too far these days.
Replying to @pitdesi
Super relevant update on the Airtable acquisition!
They spun out their AI business Hyperagent prior to the acquisition.
So the company lives on, and likely some cash with it, probably can be recapitalized.
seems like a fantastic outcome. Shed the old business and focus on AI
Vidhur Kumar retweeted
Supply chains broke my mental model of enterprise software.
I'd spent years thinking enterprise software was about helping a company optimize its own business, then I started building for supply chains and realized there isn't a single owner of the system you're trying to optimize. Decisions depend on suppliers, carriers, customers, ports, weather, tariffs, and events no one controls.
Before you can optimize a supply chain, you first need to understand how decisions ripple across the entire network, so we built Atlas to see the end-to-end picture.
Check out @vidhur2k's deep dive.
As we've scaled up AI deployments with customers over the past year, we've built up a large supply chain dataset across a multitude of use-cases, from semiconductors and GPUs to frozen foods and chemicals.
Pallet Labs was started for two reasons:
1. Advance the frontier of AI research in supply chain and provide our customers with access to the most performant models using their own data.
2. Develop a deep understanding of supply chain economics across the world.
Excited for all the work being done here!
Today, we're introducing Pallet Labs, the home for applied AI research at @trypallet.
Supply chain is a network of business contracts, technical know-how, economics, and physical operations. Working with some of the world's largest enterprises gives us a unique opportunity to study problems and proprietary datasets that don't exist in the public domain.
We've always believed that the biggest breakthroughs come from making ambitious bets years before they're obvious. One of our earliest bets was training our own models. Today, those models are live with multiple customers.
We've brought together AI researchers and economists from Ramp, Scale AI, Stanford University, Columbia University, and Duke University to tackle frontier problems and ultimately build systems that expand the world's productive capacity.
We're hiring. DM us if you'd like to join our team.
Vidhur Kumar retweeted
Today we're launching @Pallet Custom Models, giving every supply chain team true ownership of its AI: agents, models, and hardware. Customers get better outcomes at lower price.
ROI is important, but so is sovereignty. The proprietary knowledge behind every company's better service, higher margins, and lasting enterprise value should remain uniquely theirs.
Massive unlock for @TryPallet with customers across the world who care about making su e their data remains their own!
We're scaling our teams across research, engineering, and deployments. If you're interested in building with the best team in the supply chain space, join us!
Today, we're introducing @trypallet Custom Models, giving every supply chain company enterprise sovereignty over its AI.
Frontier models are powerful, but production workflows require repeated model calls, driving up costs as context and volume grow while exposing your proprietary knowledge. You pay for every token processed, not outcomes achieved.
Custom Models are trained exclusively on your operation, deployed on infrastructure you control, and never shared externally. The result is AI that's more accurate, private, and cost-efficient.















