We start with due diligence for VC and PE teams, routing only the uncertain calls to human review.
I see firsthand why extraction alone never scales: review is the bottleneck.
I also spoke on how the human brain inhibits its way to near-perfect memory and retrieval.
I built it on the Runlog AI API at Redesign Health, and spent a YC weekend building an IDE for product managers.
I built systems from scratch handling 1M+ geometries and cut user-facing latency by 70%.
Accuracy went up 15%. I learned that confidence scores without operational routing are meaningless.
We dropped third-party dependencies by building it in-house, and I added narration, large font and high contrast.
I'm founder of Runlog Atlas, and I led engineering teams at Stax and Dirac before building AI systems at Meta and Microsoft.
We start with due diligence for VC and PE teams, routing only the uncertain calls to human review.
I see firsthand why extraction alone never scales: review is the bottleneck.
I also spoke on how the human brain inhibits its way to near-perfect memory and retrieval.
I built it on the Runlog AI API at Redesign Health, and spent a YC weekend building an IDE for product managers.
I built systems from scratch handling 1M+ geometries and cut user-facing latency by 70%.
Accuracy went up 15%. I learned that confidence scores without operational routing are meaningless.
We dropped third-party dependencies by building it in-house, and I added narration, large font and high contrast.
I'm founder of Runlog Atlas, and I led engineering teams at Stax and Dirac before building AI systems at Meta and Microsoft.
We start with due diligence for VC and PE teams, routing only the uncertain calls to human review.
I see firsthand why extraction alone never scales: review is the bottleneck.
I also spoke on how the human brain inhibits its way to near-perfect memory and retrieval.
I built it on the Runlog AI API at Redesign Health, and spent a YC weekend building an IDE for product managers.
I built systems from scratch handling 1M+ geometries and cut user-facing latency by 70%.
Accuracy went up 15%. I learned that confidence scores without operational routing are meaningless.
We dropped third-party dependencies by building it in-house, and I added narration, large font and high contrast.
By using this service, you agree to the Terms of Service and Privacy Policy.