Dmitry Petrov

🛠️ Building data infra for AI/ML. Ex-Data Scientist @Microsoft. Created DVC, now DataChain. PhD in CS. Serious about data. Less serious about everything else.

[Dmitry Petrov @FullStackML]

@startupandrew @TaskletAI Wow, congrats!

[Dmitry Petrov @FullStackML]

@ajs6888 Exactly 😄 this is where the fun begins

(Glad X can translate in one click)

[Dmitry Petrov @FullStackML]

Physical AI data breaks the usual agent loop.

Tomorrow: Claude Code over raw video, metadata, and expensive compute in S3.

Live demo: 8am San Francisco / 4pm London / 5pm CET

Join us: https://t.co/gDoBGZqwDr

[Dmitry Petrov @FullStackML]

Going deeper this Thursday - with a live Claude Code demo over raw Physical AI data in S3.

8am San Francisco / 11am NY / 4pm London / 5pm CET

https://t.co/gDoBGZqwDr

[Dmitry Petrov @FullStackML]

Claude Code and Codex were built for weightless data.

Physical AI data has mass.

My talk on the other physics of agents:
https://t.co/9yn7i9n5hx

[Dmitry Petrov @FullStackML]

This made me check my own citations. Turns out DVC has 40+ citations which is more than my PhD research

Apparently, open source is a better academic strategy than academia 😄

@shcheklein @rkuprieiev @isidentical congrats, we’re academics now. Who knew?

[Ash Vardanian @ashvardanian]

You don’t have to publish papers to have academic citations or even an h-index.

I dropped out of uni years ago 🤦‍♂️, and always wanted to contribute in different ways: open-source libraries, datasets, benchmarks, and overly long blog posts. I finally created a Google Scholar profile and started linking my GitHub repositories and articles.

It turns out some academics are kind enough to cite good ideas regardless of where they come from 🫶

Scholar’s attribution machinery is hilariously basic: it split citations between GitHub and Zenodo versions, missed entire projects, and failed to connect several papers to the software or articles they explicitly reference. Nevertheless, I have already amassed a monumental h-index of 3 😂

Open-source devs, your software may already be part of the academic literature. You might even have an h-index without knowing it. Go check Google Scholar, you may be pleasantly surprised!

P.S. Geoffrey Hinton’s h-index is 192. Only 189 to go.

[Dmitry Petrov @FullStackML]

Strongly agree. Today's coding RLVR mostly targets functional correctness, not design or maintainability.

The real leap comes when RLVR adds multi-objective evaluation for design, code quality, and maintainability.

Then comes the sequel, @dexhorthy: why great software still needs humans 😄

[Dmitry Petrov @FullStackML]

@signulll My ex-teammate made the same choice. A US PhD and Big Tech experience made it much easier for him to start a company in China.

[Dmitry Petrov @FullStackML]

Coding-agent intuition breaks down on physical-world data

That’s my AI Engineer World’s Fair Online Track talk this week.

See you there! I’ll be the guy in the Villain-Con shirt 😄

[Dmitry Petrov @FullStackML]

Claude and ChatGPT are both down.

Tokenmaxing seems to work.

[Dmitry Petrov @FullStackML]

@ChiefScientist This is how we vibe coded in 2000s

[Dmitry Petrov @FullStackML]

@fantopy It’s about Data Agents over S3. Semantic Layer for Files.

[Dmitry Petrov @FullStackML]

Call for Proposals for AGNTCon + MCPCon North America 2026 is open on Sessionize and I've just submitted a session! https://t.co/bcg6jjoDdn

[Dmitry Petrov @FullStackML]

Talked about data agents beyond SQL at @pydatalondon today.

Files, video, sensors, metadata, Python workflows - all the messy stuff where real data work happens.

Thanks everyone who came by 🙌

[Dmitry Petrov @FullStackML]

@trikcode 5 —> 128 Another proof that 10x productivity isn’t a limit

[Dmitry Petrov @FullStackML]

@ArchiveExplorer Is it because Claude code, I guess?

[Dmitry Petrov @FullStackML]

https://t.co/1RTpA3zIHd

[Dmitry Petrov @FullStackML]

Excited to speak at @PyData London this Sunday!

I’ll be talking about data context for agents beyond SQL: video, sensors, files, metadata, and Python workflows.

10:15 · Hardwick Hub Conference Room

Come say hi!

[PyData @PyData]

Tomorrow, PyData London 2026 begins. If you have your ticket, we'll see you there. If you don't have yours yet — there's still time: https://t.co/SpNnxOTlar
- - -
PyData London 2026:
Convene Sancroft, St. Paul's
Tutorials: Friday, June 5-7
Talks: Saturday and Sunday, June 6-7
Keynote Speakers all 3 days
Tickets: https://t.co/SpNnxOTlar

[Dmitry Petrov @FullStackML]

@Hans365days Depends on the problem but yes, context engineering usually takes majority of the time. In data likely 90%

[Dmitry Petrov @FullStackML]

@BlackDotttt Find 5 differences 🙂

[Dmitry Petrov @FullStackML]

@abenz_mato Yes, data lineage plays an interesting role in data agents. With human it's nice to have, with agent is absolutely must-have.