Pinned Tweet
Proud to announce that my research paper on "Using medical ADT data to predict hospital readmissions" has been published in Elsevier/ScienceDirect Journal - MLWA. Feedbacks are welcome.
https://teapawt.net/t.co/XLIUr54ta6
#AI #MachineLearning #DataScience #BigData #healthcare #MLOps
🎙️ DTM podcast: Join me & @raodyboy, co-founder of @SpecialeInvest, as we dive deep into VC firm structure, essential skills for VCs, networking's power, & the future of VC & deep tech startups! 🚀
Listen or read the summary here - bit.ly/dtm-as
#deeptech #VC #india
#DTM #Podcast: The Inspirational Story of Building a Healthcare AI Startup From 0 to 1 in India and Taking it to the World - Dr. @geethamhp, founder of @niramaianalytix.
Spotify link: bit.ly/48vhLii
#healthtech #startup #india
Pronojit Saha retweeted
Whenever you see such scenes: always ask-- where is your tax money going?
While the common people play musical chairs to get a seat. And, continue to pay unfair taxes.
The Rich benefit from easy debt. And, corporate loans waivers (which have now reached 10 Lakh crores in the last 5 years).
#DTM Podcast: Unlock the insights behind @piventures success: Founding story, investment focus, deep tech themes for the future, and more. Learn from one of the best minds in Deep Tech @manish_saarthi.
Listen or read here - bit.ly/dtm-mp
#AI #Angels #podcast
#DTM Podcast🎙️Uncover insights into @robmay's 20-year content success, the rise of synthetic users, horizontal LLMs & future of OpenAI, achieving "Agent-Market Fit", and AI's impact on competitive advantage.
Listen or read here - bit.ly/dtm-rb
#AI #Podcast
#DTM Podcast: Avnish is driving startup success at
@Accenture Ventures. With 30yrs of exp, he shares his strategies for jumping career S-curves, forming partnerships with large corporations, fostering innovation & capitalizing on GenAI.
Read or listen- bit.ly/dtm-aa
#DTM #Podcast🚀 Sonal, VP at @3one4Capital is driving exponential growth by partnering with generational companies & transforming markets. Join us as we explore startup success, deep tech, investment strategies, and more!🌟
Listen or read summary here - bit.ly/dtm-s3
Pronojit Saha retweeted
In addition to prompting LLMs, many developers are now also experimenting with fine-tuning. I describe in The Batch how to choose from the growing menu of options for building applications with LLMs: Prompting, few-shot, fine-tuning, pre-training. deeplearning.ai/the-batch/is… – at Loyola, CA
Pronojit Saha retweeted
This seems to be written in good faith, so I'll provide (one) Indian VC's perspective.
The most salient fact about India venture is the extreme scarcity of late-stage capital providers*. You can count on the fingers of one hand funds who can write a $30M cheque on their own. 1/
One stark difference when I speak with US VCs vs Indian VCs.
Indian funds seem to be looking for signals of the market being not that competitive, as they hold that most markets are winner-takes-all.
US funds seem to be more focused on the merit of the company without much heed to the competitive landscape as most markets see multiple winners.
If anything, having competitors/comparable companies is often a validation of the space/problem being large enough.
Pronojit Saha retweeted
Three ingredients to a great B2B startup idea:
1. You solve a problem that is important to people
2. Existing solutions are doing a subpar job of solving the problem
3. You’re excited about solving this problem
More below.
🎙️#DTM New Episode🎙️
Join me as I chat with Ankit, founder of ZedBee (incubated at @iitm_respark), an energy management startup using IoT & AI/ML.
Get insights on pre-seed funding, idea validation, & building startup moats!
Listen or read summary here - bit.ly/dtm-az
Pronojit Saha retweeted
Use cases for encoder LLMs (classification) & decoder LLMs (chatbots) are obvious.
Seq2seq tasks like translation, where it makes sense to have access to the whole input, is where it gets interesting.
Re encoder-decoder LLMs (eg T5): Is there still a benefit of using an encoder?
Pronojit Saha retweeted
2nd SWP (Systematic Withdrawal Plan) thread on conservative withdrawal.
Lumpsum: 1 Cr & 50 Lakh
Time Frame: 12yrs
Withdrawal %: 6%
Assumption: have own house, kids are settled & they are moving partial savings in SWP.
I am picking funds randomly. No Discrimination.
Lets go⤵️
Pronojit Saha retweeted
Love these new summaries @lennysan is doing of podcast episodes.
Here’s one of mine: lennyspodcast.com/blog/defin…
Pronojit Saha retweeted
Finally got a chance to spend some time with the 77-page Llama 2 paper (arxiv.org/abs/2307.09288). Here are some takeaways at a glance.
Appreciate that someone finally did a comprehensive supervised finetuning vs RLHF evaluation! (Lower right)
🎉 We've hit the big 50!
Join us for this special 50th episode of DTM podcast with the OG of Indian Deep-Tech, @arpiit from @BlumeVentures as we explore the history, challenges, and future of Deep-Tech in India.
Listen or read the summary here - bit.ly/dtm-ab
Thanks to all listeners, guests, and the thriving Indian deep-tech community for your continued love and encouragement.
@sajithpai: Thanks for your support on this episode!
#deeptech #cleantech #biotech #spacetech #ai #machinelearning #datascience #India #VC
3
282
Pronojit Saha retweeted
If you are a SaaS startup from India with aspirations to expand to the US, you would do well to read this.
Exceptional post from @awadehra who is emerging as a leading US GTM / PMF Whisperer for Desi SaaS startups.
Key highlights below.
blume.vc/commentaries/so-you…
Pronojit Saha retweeted
"Python vs R" is now finally solved 😅
Hi #EconTwitter!
Exciting news🎉:
@robtibshirani-@daniela_witten-James-Hastie-Taylor have released the #python version of the influential 𝐄𝐥𝐞𝐦𝐞𝐧𝐭𝐬 𝐨𝐟 𝐒𝐭𝐚𝐭𝐢𝐬𝐭𝐢𝐜𝐚𝐥 𝐋𝐞𝐚𝐫𝐧𝐢𝐧𝐠!
Course, datasets, notebook (and more) included. 📚📊
Don't miss out!
Pronojit Saha retweeted
Good thread deconstructing the Pharmeasy downround - tldr is that the biz model had weak unit econ baked into it (thanks to pharma pricing regulations), which no amount of scale could overcome.
Which brings me to 2 points.














