Pinned Tweet
Wouldn’t it be nice to just ask what you need from your data?
You can now.
Ask questions in plain English, using your AI tools (Grok, Claude Code, etc.) on your machine, interactive response time. Direct from object storage.
AirMettle Select - on Azure.
(US East 2, contact us for timing for other regions / clouds)
1
1
330
Tariffs would have been an easy sell and POPULAR if they made ~ 90% of them from a given quarter get refunded to resident citizens (not those living overseas)
it would make it cost neutral (likely net-positive) to most people for what they buy.
13
This video explains how AirMettle Select achieve such low-latency analytics on your raw data, which remains under your control.
1
78
Donpaul Stephens retweeted
Today, Carnegie Mellon University and Ken Griffin, founder and CEO of Citadel and founder of Griffin Catalyst, announced a historic $3 billion gift to elevate Carnegie Mellon in Pittsburgh and launch Carnegie Mellon University Miami. #CMUMiami
Learn more: cmu.edu/news/stories/archive…
AirMettle’s monitoring solution is in early access - reach out if interested.
Drop-in alternative for Prometheus (OpenTelemetry Collector and PromQL support for Grafana & more)
With an MCP server for natural language queries via LLMs (Claude & Grok)
More importantly than faster speed and more efficient storage (lower cost). significantly more precise than Prometheus histograms while storing significantly fewer series to hold - and has native support for monitoring correlations (2D+)
1
2
222
AirMettle Select is now live on the Azure Marketplace.
Low-latency SQL on large logs direct from your Blob Storage (CSV, JSON, JSONL, gzip).
10x to over 100x faster than other on-demand solutions. Not a warehouse. You keep your data.
• Leave data in native format
• GZIP lowers cost without impacting query performance
• Feed filtered results into Fabric
airmettle.com/select
#Azure #SQL #BigData
8
319
AirStat is now live on the Microsoft Azure Marketplace.
Point it at numerical data already in your Azure Blob Storage, ADLS Gen2, or OneLake. AirStat turns enormous volumes of data into compact statistical fingerprints — typically 3+ orders of magnitude smaller — that preserve exact statistics and can be rapidly merged and rolled up across sensors, across space, and/or across time.
From these fingerprints you can extract true percentiles, full-resolution CDFs, top-k outliers, joint distributions, and multi-dimensional views — all exactly, in milliseconds, without rescanning the raw data!
Your data never leaves your account. Nothing to deploy. Fully compatible with Microsoft Fabric and OneLake.
Get started: marketplace.microsoft.com/en…
We turned the AirTree engine into a fully managed service.
It builds compact statistical fingerprints from numerical data already in your cloud storage. These fingerprints enable interactive exploration of the characteristics of massive datasets — delivering exact answers in milliseconds.
You explore the fingerprints, not the raw data. This makes true interactive exploratory analytics practical at scale.
Unlike many sketch-based approaches that are embedded inside larger platforms, this works directly against your files and is easily scriptable into existing pipelines and workflows.
Our website has live interactive examples on real public datasets. You can drag to query distributions, explore multi-dimensional views, and zoom from a high-level overview down to fine detail.
Come try them: airmettle.com/airstat
What if you could turn enormous amounts of numerical data into a compact statistical fingerprint?
These fingerprints are typically 3+ orders of magnitude smaller than the original data. They preserve exact statistics and can be merged losslessly — across sensors, across shards, or across time.
From a single fingerprint (or a collection of them) you can extract:
• True percentiles and full-resolution CDFs
• Real top-k outliers
• Joint distributions and multi-dimensional grids / heatmaps
All exactly. With zero approximation error. In milliseconds. Without rescanning the raw data.
That’s AirTree.
While this may seem almost impossibly compact, you can try the engine yourself here:
github.com/AirMettle/AirTree
Source-available for non-commercial use. Commercial licensing available.
Donpaul Stephens retweeted
Crunchyroll is bringing the legendary cyberpunk icon AKIRA back to theatres in stunning 4K and IMAX! Relive the film on the big screen this September 🔥
Read more: got.cr/AKIRA4K-x
Donpaul Stephens retweeted
Win or lose, respect to @elonmusk for being the man in the arena
My team and I are working on a new visual radar system designed to help airports detect things that today often slip through the cracks in the surrounding area; drones, birds, and other small objects that can cause delays or go-arounds. We started this project within the US Space Force’s SDA TAP Lab.
Our goal is simple: Make runway safety smarter, more reliable, and dramatically more affordable than traditional surveillance systems.
Before we go any further, I want to learn directly from the people who know this world best. If you work in airport operations, safety, airfield management, or anything similar, your perspective would be incredibly valuable. -- please reply and follow (so we can DM to coordinate a time to speak)
9
904
Next week, AirMettle will be demonstrating our solution to
Find anything flying
Using standard cameras (no moving parts) - backed by our VelocInsight analytics which is commercially licensed for use by a National Lab and is in early access trials in Azure.
NIH spearheads health innovations with ~$1.3B in FY23 across diverse divisions spanning cancer, infectious diseases, aging, and more—from medical devices to software diagnostic tools to nascent drugs, and beyond. The NIH funds a broad Phase I pool for concept proof but is highly selective for Phase II due to biomed's capital intensity and need for rigorous validation before patient impact.
seed.nih.gov/aboutseed
1
39
NASA provides ~$250M in FY23, with programs for NASA-specific tech (as primary customer) and commercially focused Ignite topics.
nasa.gov/sbir-sttr
1
30







