iiinova — where your digital self keeps evolving

Diptanu Choudhury @diptanu.bsky.social · 10/04/2026

If you are building agents, would love to have a conversation. While many companies in this space focuses on developer experience, we are also focused on performance and consistency beyond the DX!

130

Diptanu Choudhury @diptanu.bsky.social · 10/04/2026

I need to post here more often! We have been working on sandboxes for agents @tensorlake.ai and have built the fastest sandbox infrastructure that is out there. Fastest file system, consistent sub second startup times and stateful resume and suspend of sandboxes for durable agents!

130

Reposted by Diptanu Choudhury

David Calavera @calavera.dev · 15/09/2025

Job update: a couple of weeks ago, I joined @tensorlake.ai full time. I’m having a lot of fun building the product with @diptanu.bsky.social and the rest of this wonderful team. We have a few open positions if you’d like to work with us: www.linkedin.com/jobs/search/...

194

Diptanu Choudhury @diptanu.bsky.social · 11/09/2025

Some more color on this. What does it mean to help models solve interesting problems with MCP? MCP is the easiest way to give models access to the gazzilion APIs enterprises have already built. yes, it’s a security nightmare but nevertheless it helps teams with the Day 1 problems of 🚢 MVPs

150

Diptanu Choudhury @diptanu.bsky.social · 11/09/2025

@chris.blue your post is spot on about the technical merits(or lack of) of MCP. Just sharing perspectives of people I have talked to :)

010

Diptanu Choudhury @diptanu.bsky.social · 11/09/2025

I was a skeptic until I heard banks and hedge funds using MCP to enable models solve interesting business problems. And yes OpenAPI + tool calls would have just worked fine but developers like frameworks and abstractions for building the first MVP. MCP being more prescriptive thank OpenAPI helps.

130

Diptanu Choudhury @diptanu.bsky.social · 18/07/2025

One of our customers told me today the biggest lift for structured extraction with @tensorlake is that their engineering team can now tweak the schema they want to extract from documents every week as they evolve their insurance platform. These little things make AI find roots in enterprises.

020

Reposted by Diptanu Choudhury

Tensorlake @tensorlake.ai · 15/05/2025

Announcing Tensorlake Cloud Up-leveling Document Ingestion and Workflows for building agentic applications and complex business workflows. www.youtube.com/shorts/OCv-...

youtube.com

🚀 Tensorlake Cloud is Here: Document Ingestion and Orchestration for AI Workflows

Tired of brittle pipelines and unreliable document parsing? Tensorlake is the only platform built for mission-critical document ingestion and orchestration i...

173

Diptanu Choudhury @diptanu.bsky.social · 07/04/2025

This would be a problem if you used the data in an agentic application for a real estate business. This is why Table Structure Understanding models are still the most reliable way to parse tables in business critical applications.

020

Diptanu Choudhury @diptanu.bsky.social · 07/04/2025

Here's an instance where Gemini Flash 2 misinterpreted a table, sourced from a residential property's mold report. It mistakenly attributed mold data from outside the property to a bedroom, as it failed to parse some rows from the third column

330

Diptanu Choudhury @diptanu.bsky.social · 24/03/2025

The emergence of browser agents makes me feel we are lacking a google for APIs. Instead developers are making LLMs search for things, and turning websites into APIs by making LLMs click on the UI like humans. This approach is going to take a long time to get right, and wasteful for a lot of reasons

130

Diptanu Choudhury @diptanu.bsky.social · 19/03/2025

It works pretty well! I will create an open source repo, with some code which ingests all these PDFs and creates datasets for people to work :)

020

Diptanu Choudhury @diptanu.bsky.social · 19/03/2025

. @tensorlake.bsky.social is pretty good! Here is our Document Ingestion API working on a scanned page from the recently released Kennedy assassination classified reports. Head over to tensorlake.ai for more! (The product is still in early preview)

061

Diptanu Choudhury @diptanu.bsky.social · 31/01/2025

Test prompts instantly with our intuitive JSON schema editor and extract complex tables, figures, and charts with ease. Once your schema is set, run our API at scale—tens of thousands of times per day—ingesting new documents and writing structured data directly into your database or warehouse.

040

Diptanu Choudhury @diptanu.bsky.social · 31/01/2025

Structured Extraction is essential for AI engineering teams, we are now making it faster and more reliable than ever, whether you're turning PDFs, invoices, or reports into structured data. Here is a sneak peak into our Structured Extraction engine.

1122

Diptanu Choudhury @diptanu.bsky.social · 27/12/2024

I don’t live in New York but will come out just for that :)

070

Diptanu Choudhury @diptanu.bsky.social · 24/12/2024

Does Dagster have a way of packaging workflows into artifacts along with dependencies? Looks like in your screenshot you are importing requests. Do you build some docker images out of band and make sure the workflow functions run on those images?

020

Diptanu Choudhury @diptanu.bsky.social · 19/12/2024

This means we won’t need a dynamodb table anymore right?

140

Diptanu Choudhury @diptanu.bsky.social · 18/12/2024

Meant creating an artifact which includes the system and python dependencies. I have seen some tools copying code into a container with system dependencies, which others pickling the code and putting it into a container.

200

Diptanu Choudhury @diptanu.bsky.social · 17/12/2024

Python Folks - which data/workflow engine has the best developer experience for packaging code? We have looked into - Modal, Beam, Airflow, Flyte, AWS Lambda, Prefect, Dagster and Spark. Haven’t seen any approach which is fast, reliable and intuitive.

6102

Diptanu Choudhury @diptanu.bsky.social · 14/12/2024

Taking a break for 10 days for the first time since December last year! January is going to be great and you will hear about @tensorlake.bsky.social more often :)

060

Diptanu Choudhury @diptanu.bsky.social · 09/12/2024

We have been using O1 or Sonnet to solve a problem to understand the upper bounds of what models are capable of, and falling back to our internal models or open source models for economy and security. Been working pretty well, is this a common workflow?

041

Diptanu Choudhury @diptanu.bsky.social · 04/12/2024

I know man. I poured my heart out on nomad back in the day. It was a ton of hard work. Thanks for appreciating :)

110

Diptanu Choudhury @diptanu.bsky.social · 04/12/2024

At the Hasicorp ReInvent party, no mention of Nomad and Consul 😭

150

Diptanu Choudhury @diptanu.bsky.social · 03/12/2024

Landed in Vegas for reinvent! Say hello if you are around, would love to chat :)

020

Diptanu Choudhury @diptanu.bsky.social · 02/12/2024

Turned on Apple Intelligence this morning. We are a long way from having a personal assistant on the iPhone! It wish it summarized all unread from Slack, Gmail, WhatsApp and messages and came up with a list of things I needed to respond :)

130

Diptanu Choudhury @diptanu.bsky.social · 01/12/2024

This is why you see Anthropic and OpenAI go up the stack and integrating document sources and building agentic frameworks and building a better chat/workspace product.

000

Diptanu Choudhury @diptanu.bsky.social · 01/12/2024

Alibaba has done an amazing job with open source models. At this point, the difference between @Alibaba_Qwen and closed vendors is just the product on top of models.

110

Diptanu Choudhury @diptanu.bsky.social · 01/12/2024

Assuming you have a great document segmentation model (which is a big assumption) it's going to do the job better than anything else out there.

020

Diptanu Choudhury @diptanu.bsky.social · 01/12/2024

Qwen2VL 72B is just better than every other closed and open source vision model for document understanding. Like every other vision model, it's still incapable of retaining every single ground truth on dense documents.

130

Diptanu Choudhury @diptanu.bsky.social · 30/11/2024

Throwing the kitchen sink at a small problem. Whenever I work on an Applied AI problem I work with unconstrained compute to see if we can solve a business problem if money was not a constraint. If there is enough value in solving the problem, the economy of scale can kick in later.

030

Diptanu Choudhury @diptanu.bsky.social · 29/11/2024

Does NVIDIA have a 2 x H100 SKU or cloud vendors are slicing up 8 x H100 machines into 4 VMs?

031

Diptanu Choudhury @diptanu.bsky.social · 28/11/2024

If you are processing unstructured data(document, audio, raw text, etc) - you really want to write Python or Typescript with the ability to load data from blob stores, and have a GPU attached to the runtime to be able to leverage AI models. SQL is a big impedance mismatch for these use cases.

030

Diptanu Choudhury @diptanu.bsky.social · 28/11/2024

Building a solid compute engine is time and capital intensive - Probably a big reason why we see execution engines use SQL as a front end. They can re-use some parts of the planner and the DSL. But then they make the trade off of pushing SQL in domains where it doesn’t make sense.

120

Diptanu Choudhury @diptanu.bsky.social · 28/11/2024

It depends on what lens you use to see things. Langchain/Graph is great for someone learning programming and AI. At sufficient scale and complexity, developers often invent a homegrown framework. What remains from the original prototypes are the shim between models and databases.

010

Diptanu Choudhury @diptanu.bsky.social · 25/11/2024

I feel they have done a great job tbh. So many people has had their first intro to AI through their docs. The speed at which they develop often is at the odds with more permanent long term solutions but it gets people off to a decent start. Reminds me of J2EE from back in the day.

100

Diptanu Choudhury @diptanu.bsky.social · 25/11/2024

Feels like at this point Bitcoin is inevitable. Both sides of the aisle seems to be supporting crypto.

100

Diptanu Choudhury @diptanu.bsky.social · 24/11/2024

Would love to be added! Founder of Tensorlake :) - we are building a serverless platform for processing unstructured data!

010

Diptanu Choudhury @diptanu.bsky.social · 24/11/2024

Well that’s orchestration, but completely serverless so you don’t have to deal with running an orchestrator :)

020

Diptanu Choudhury @diptanu.bsky.social · 23/11/2024

Nice. We will give them a go! We built the control plane to be able to leverage distributed data plane, so we can use pretty much any provider under the sun. Need disk, connectivity to the control plane APIs and GPUs :)

000

Diptanu Choudhury @diptanu.bsky.social · 23/11/2024

How’s the pricing compared to lambda labs?

000

Diptanu Choudhury @diptanu.bsky.social · 23/11/2024

Done Adam! Please share for visibility :) You guys are doing great work at Prefect :)

010

Diptanu Choudhury @diptanu.bsky.social · 23/11/2024

Have you used them? I haven’t used DO in a really long time, how’s the service these days?

200

Diptanu Choudhury @diptanu.bsky.social · 23/11/2024

Done! Please share it for visibility :)

110

Diptanu Choudhury @diptanu.bsky.social · 20/11/2024

Sounds great! Sorry I didn’t realize there is literally no description on our account! :)

100

Diptanu Choudhury @diptanu.bsky.social · 20/11/2024

Not at the moment, we should write a blog about this soon!

110