Diptanu Choudhury
@diptanu.bsky.social
CEO @tensorlake.bsky.social Past - AI Infrastructure at Facebook, LinkedIN, Hashicorp, Netflix
Pinned
Diptanu Choudhury
@diptanu.bsky.social
· Nov 17
Applied AI Folks!
Join the conversation
buff.ly
Just created a starter pack to get a steady stream of Applied AI related conversations! Think RAG in Enterprises, Structured Extraction from documents, AI in consumer apps and Agents!
I will keep adding folks to this list! Who else should I add?!
https://buff.ly/3ObxkCD
I will keep adding folks to this list! Who else should I add?!
https://buff.ly/3ObxkCD
Reposted by Diptanu Choudhury
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/...
We have a few open positions if you’d like to work with us: www.linkedin.com/jobs/search/...
September 15, 2025 at 7:29 PM
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/...
We have a few open positions if you’d like to work with us: www.linkedin.com/jobs/search/...
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
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
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.
New post! MCP seems like a lot of hype without much substance. Seems like some developers are catching on.
September 11, 2025 at 5:11 PM
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
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
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.
New post! MCP seems like a lot of hype without much substance. Seems like some developers are catching on.
This MCP Server Could Have Been a JSON File
There's a lot of buzz around MCP. I'm not convinced it needs to exist.
materializedview.io
September 11, 2025 at 4:30 PM
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.
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.
These little things make AI find roots in enterprises.
July 18, 2025 at 4:17 AM
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.
These little things make AI find roots in enterprises.
Reposted by Diptanu Choudhury
Announcing Tensorlake Cloud
Up-leveling Document Ingestion and Workflows for building agentic applications and complex business workflows.
www.youtube.com/shorts/OCv-...
Up-leveling Document Ingestion and Workflows for building agentic applications and complex business workflows.
www.youtube.com/shorts/OCv-...
🚀 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...
www.youtube.com
May 15, 2025 at 4:06 PM
Announcing Tensorlake Cloud
Up-leveling Document Ingestion and Workflows for building agentic applications and complex business workflows.
www.youtube.com/shorts/OCv-...
Up-leveling Document Ingestion and Workflows for building agentic applications and complex business workflows.
www.youtube.com/shorts/OCv-...
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
April 7, 2025 at 4:36 AM
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
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
March 24, 2025 at 3:45 PM
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
. @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)
Head over to tensorlake.ai for more! (The product is still in early preview)
March 19, 2025 at 5:10 AM
. @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)
Head over to tensorlake.ai for more! (The product is still in early preview)
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.
Here is a sneak peak into our Structured Extraction engine.
January 31, 2025 at 7:08 PM
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.
Here is a sneak peak into our Structured Extraction engine.
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.
December 17, 2024 at 4:09 PM
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.
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 :)
December 14, 2024 at 8:05 PM
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 :)
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?
December 9, 2024 at 8:12 PM
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?
At the Hasicorp ReInvent party, no mention of Nomad and Consul 😭
December 4, 2024 at 3:23 AM
At the Hasicorp ReInvent party, no mention of Nomad and Consul 😭
Landed in Vegas for reinvent! Say hello if you are around, would love to chat :)
December 3, 2024 at 12:02 AM
Landed in Vegas for reinvent! Say hello if you are around, would love to chat :)
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 :)
It wish it summarized all unread from Slack, Gmail, WhatsApp and messages and came up with a list of things I needed to respond :)
December 2, 2024 at 3:35 PM
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 :)
It wish it summarized all unread from Slack, Gmail, WhatsApp and messages and came up with a list of things I needed to respond :)
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.
December 1, 2024 at 10:58 PM
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.
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.
Like every other vision model, it's still incapable of retaining every single ground truth on dense documents.
December 1, 2024 at 9:41 PM
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.
Like every other vision model, it's still incapable of retaining every single ground truth on dense documents.
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.
November 30, 2024 at 9:21 PM
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.
Does NVIDIA have a 2 x H100 SKU or cloud vendors are slicing up 8 x H100 machines into 4 VMs?
November 29, 2024 at 10:59 PM
Does NVIDIA have a 2 x H100 SKU or cloud vendors are slicing up 8 x H100 machines into 4 VMs?
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.
But then they make the trade off of pushing SQL in domains where it doesn’t make sense.
November 28, 2024 at 9:58 PM
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.
But then they make the trade off of pushing SQL in domains where it doesn’t make sense.
I love that DMs are open to chat with people on BSKY by default! Had some great conversations with folks today!
Please DM if you are working on anything related to unstructured data, LLMs and Document Understanding!
Please DM if you are working on anything related to unstructured data, LLMs and Document Understanding!
November 19, 2024 at 11:47 PM
I love that DMs are open to chat with people on BSKY by default! Had some great conversations with folks today!
Please DM if you are working on anything related to unstructured data, LLMs and Document Understanding!
Please DM if you are working on anything related to unstructured data, LLMs and Document Understanding!
Just created a starter pack to get a steady stream of Applied AI related conversations! Think RAG in Enterprises, Structured Extraction from documents, AI in consumer apps and Agents!
I will keep adding folks to this list! Who else should I add?!
https://buff.ly/3ObxkCD
I will keep adding folks to this list! Who else should I add?!
https://buff.ly/3ObxkCD
Applied AI Folks!
Join the conversation
buff.ly
November 17, 2024 at 8:40 AM
Just created a starter pack to get a steady stream of Applied AI related conversations! Think RAG in Enterprises, Structured Extraction from documents, AI in consumer apps and Agents!
I will keep adding folks to this list! Who else should I add?!
https://buff.ly/3ObxkCD
I will keep adding folks to this list! Who else should I add?!
https://buff.ly/3ObxkCD
With AWS's GPU pricing I wonder if any early stage companies which can't make multi-year commitments to getting X amount of GPU hardware need to have a multi-cloud strategy from Day 1.
November 12, 2024 at 6:50 AM
With AWS's GPU pricing I wonder if any early stage companies which can't make multi-year commitments to getting X amount of GPU hardware need to have a multi-cloud strategy from Day 1.
Looking forward to this! We just started using SlateDB in @tensorlake.bsky.social’s compute engine, looking forward to not needing dynamodb for CAS. One less moving part!
@chris.blue when is the new release coming? :)
@chris.blue when is the new release coming? :)
And it's done. Thank you so much @benesch.bsky.social for doing this. Now SlateDB will be able to have CAS for all major object stores. 🔥 It just works.
object_store: Conditional put and rename_if_not_exist on S3 · Issue #6285 · apache/arrow-rs
Is your feature request related to a problem or challenge? Please describe what you are trying to do. This is now supported on S3: https://aws.amazon.com/about-aws/whats-new/2024/08/amazon-s3-condi...
github.com
November 9, 2024 at 4:07 PM
Looking forward to this! We just started using SlateDB in @tensorlake.bsky.social’s compute engine, looking forward to not needing dynamodb for CAS. One less moving part!
@chris.blue when is the new release coming? :)
@chris.blue when is the new release coming? :)
The OCI ecosystem for building and distributing containers have improved SOOO much in the last 4-5 years! It's mind blowing.
Cold Startups for large containers are almost not a problem anymore if you have enough network bandwidth and some CPU to spare.
Cold Startups for large containers are almost not a problem anymore if you have enough network bandwidth and some CPU to spare.
November 7, 2024 at 1:38 AM
The OCI ecosystem for building and distributing containers have improved SOOO much in the last 4-5 years! It's mind blowing.
Cold Startups for large containers are almost not a problem anymore if you have enough network bandwidth and some CPU to spare.
Cold Startups for large containers are almost not a problem anymore if you have enough network bandwidth and some CPU to spare.