geek, scribe, coffee snob, and wanna-be cyclist. Contributor to Apache Spark and Delta Lake maintainer. Developer Relations at Databricks (opinions r my own)
Seattle
Joined May 2008
- Tweets19.1K
- Following1.9K
- Followers3K
- Likes27.2K
dennylee retweeted
Just to prove we shot it practically🥲
No AI, shot in-camera.
Big thanks to @umoco2020 for providing the robotic arm!
Check out the full review to see the final shots➡️xitter.cf/mediastorm_en/status/2…
dennylee retweeted
Great work on fast long-context prefill. TokenGate is a trainable module that works with existing models to make long-context inference much more efficient.
dennylee retweeted
📣 Join us Nov 4: Graph Reasoning for AI Agents.
Danfeng Xu, CTO of @puppyquery, will walk through Omnigent + PuppyGraph: graph reasoning on your existing tables, in real time, without moving a row. Skip the ETL. Keep unified governance.
Register 👇
luma.com/omnigent-bnib
#Omnigent #PuppyGraph
dennylee retweeted
Databricks and @nvidia are expanding our collaboration so enterprises can customize open models on their own data and take them all the way to production.
With AI Runtime, developers can use NVIDIA NeMo, NGC containers and new Nemotron cookbooks to train and fine-tune models on serverless NVIDIA GPUs alongside governed enterprise data. Nemotron 3.5 lightning and Super VL are also ready to load, adapt and serve on Databricks from day one.
Customize with governed enterprise data. Serve in production. All on Databricks.
databricks.com/blog/databric…
dennylee retweeted
Incredible work by @veronica3207 and team!
Introducing MemCurrent, a persistent memory architecture designed to help production agents stop starting from zero.
MemCurrent combines episodic memory with derived memory across three access patterns:
• Plugins to log and load context automatically
• Tools the agent calls on demand
• An offline “dreaming” pipeline that consolidates memory in the background
Agents carry forward user preferences, prior corrections, metric definitions and successful workflows instead of rediscovering them every session.
On our internal benchmark built from real enterprise data, MemCurrent triples the agent's pass rate while using 31% fewer tokens and running 26% faster. It's built into Agent Bricks' Memory offering and already powers memory in Genie.
Read the research: databricks.com/blog/memcurre…
dennylee retweeted
Databricks AI Runtime (AIR) is now GA, offering an easy and super-fast way to launch GPU workloads. Unlike provisioning your own GPU servers, you get GPUs with your code in seconds, with easy access to all your data. This is especially relevant now that agents can automate much of AI training, and want to scale up/down experiments quickly. Read how companies like Robinhood, Superhuman, Factset and Merck have been using AIR to accelerate AI training: databricks.com/blog/ai-runti…
dennylee retweeted
Great conversation with @BrendanFoody talking about the main bottleneck to enterprise adoption of AI and AGI: knowing whether AI is doing what you want it to. The hard part is creating a measurable specification for qualitative tasks in the real data context, i.e., an eval.
We’ve automated a very low percentage of enterprise tasks with AI.
That’s @jefrankle’s estimate, and he thinks the models are already smart enough to do significantly more.
Jonathan is the Chief AI Scientist at Databricks, co-founder of MosaicML, and my latest guest on The Frontier.
We talked about why evals are the biggest bottleneck to enterprise AI adoption. The intelligence is increasingly there, but enterprises need to figure out how to define what they want agents to do and measure whether they’re doing it well.
Full episode below 👇
dennylee retweeted
Databricks is investing in New York for the long haul. Very excited to continue growing our AI research lab in NYC, now one of the largest in the city!
New York is becoming one of Databricks’ biggest AI hubs. 🗽
We’re expanding at 5 Bryant Park to ~120,000 square feet across five floors, including a dedicated Engineering and Research floor.
With teams building technologies behind Genie, Unity Gateway and CustomerLake, our momentum in NYC is only growing.
dennylee retweeted
The energy has been electric at #DataAIWorldTour 2026, and we’re not even halfway through!
Next up:
🌬️ Thu, Oct 29 — Chicago
🇩🇪 Wed, Nov 4 — Munich
🍑 Tue, Nov 10 — Atlanta
🇳🇱 Tue, Nov 10 — Amsterdam
🇬🇧 Wed, Nov 11 — London
Across each stop, learn how organizations are grounding AI in enterprise data and business context, governing data and AI usage, and putting apps, agents and analytics into production.
Find a stop near you: databricks.com/dataaisummit/…
dennylee retweeted
Can a single database architecture satisfy both vibe-coded and large scale enterprise apps?
The same approach makes small database ultra cheap to run and branch can also make a large one faster and easier to operate. We call this architecture Lakebase.
dennylee retweeted
This is Tommy. Tommy had a question about my allegiances. Listen to the end to hear about his.
dennylee retweeted
Lakebase architecture has many benefits. One of which is making every dba operation instant.
This shift such operations left from dba to a developer and all the way to an agent.
Restoring a 100 TB database with Lakebase Postgres takes seconds.
How?
A branch-based restore is a metadata operation. It creates a branch at a point in time, referencing immutable history that already lives in decoupled storage. No copying data, no replaying WAL. Recovery is near-instant regardless of database size, and simple enough for agents to use in branching and undo workflows.
databricks.com/blog/lakebase…
dennylee retweeted
Congrats to @reflection_ai on open-sourcing Beam! 🙌
Omnigent is adding Mirror as a built-in ACP harness so it can run alongside the other harnesses in your workspace.
More open models, more choice for agent builders.
#Omnigent #OpenSource
Beam is now available across platforms, with open weights under the Apache 2.0 license - in collaboration with an ecosystem of 40+ partners.
We're releasing:
- Weights, tokenizer, and reference implementation
- BF16, FP8 and NVFP4 weights and a DFlash2 drafter
- The open source harness we built and trained Beam on: Mirror CLI
Weights and model card on Hugging Face: huggingface.co/reflection
dennylee retweeted
Cool results making auto-research / AI-based optimization better by having agents forget! Turns out that just running an agent for a long time can lock it into small steps in the same direction.
dennylee retweeted
All Neon databases now support live queries with Neon Realtime! This was much easier to build with the disaggregated architecture of lakebase, and enables so many new user experiences.
As a longtime @neondatabase user, there is one feature that I have always wanted.
Well...
dennylee retweeted
Now in public beta, Claude Platform can send OpenTelemetry traces from the Messages API, Claude Managed Agents, and the Claude SDKs to your observability stack.
dennylee retweeted
Game predictions. Player comparisons. Live Excitement Scores. Vibe Checks. FOX Sports AI delivers all of it to fans during major live sporting events, built on Databricks Lakebase.
Today we’re excited to announce a multi-year strategic partnership with Fox Corporation, making Databricks FOX's platform for data, AI apps, and agents. Sports AI is just the start. FOX is also using Lakebase, Genie Code, Unity Gateway, and Genie Ontology across streaming, content recommendations, and more.
Read the full story: databricks.com/company/newsr…
