Sarang Shelke

Seasoned semiconductor tech expert with 25 years in IP development and customer management. Building top-tier teams for tackling complex system challenges

[STPI Pune @STPIPune]

STPI-Pune is organizing a workshop for startups with renowned expert Mr. Chris Rowen, VP Engineering, Collaboration AI at Cisco focusing on key aspects such as product–market fit, customer access, funding ideas & growth strategies etc.

#STPIINDIA #Innovation
@arvindtw @stpiindia

[Sarang Shelke @sarangshelke]

@authoramish Great discussion. A small correction - Mein Samay Hun - was Harish Bhimani ji not Raza Murad ji

[Sarang Shelke @sarangshelke]

@docbhooshan Can’t have better precision missile like response than this 😂

[GDP @bookwormengr]

Writing this as an Indian who works on AI in leadership role for one the largest companies in the world (though strictly my personal opinion, but based on verifiable data).

You heard it first here:
—————————-
First some more shocks:
You heard DeepSeek.
Wait till you hear about Qwen (Alibaba), MiniMax, Kimi, DuoBao (ByteDance) all from China.
Within China, DeepSeek is not unique and their competition is close behind (not far behind).
IMHO, China has 10 labs comparable to OpenAI/Anthropic and another 50 tier 2 labs.
The world will discover them in coming weeks in awe and shock.

AI is not hard (I am not high)
————————————
Ignore Sam Altman.
Many teams that built foundation models are below 50 persons (e.g. Mixtral).
In AI, LLM science part is actually quite easy.
All these models are “Transformer Decoder only models”, an architecture that was invented in late 2017.
There are improvements since then (flash attention, ROPE, MOE, PPO/DPO/GRPO), but they are relatively minor, open source and easy to implement.

Since building foundation models is easy and Nvidia is there to help you (if not directly, then by sharing their software like “Megatron” that is assembly line to build AI models) there are so many foundation models built by Chinese labs as well as global labs.

It is machines that learn by themselves…if you give them data & compute. This is unlike writing operating system or database software. Also, everyone trains on same data: internet archives, books, github code for the first stage called “pre-training”.

What is part is hard then?
———————————-
It is the parallel & distributed computing to run AI training jobs across thousands of GPUs that is hard. DeepSeek did lot of innovation here to save on “flops” and network calls. They used an innovative architecture called Mixture of Experts and a new approach called GRPO. with verifiable rewards both of which are in open domain through 2024.

Also, there is lot of data curation needed particularly for “post training”
to teach model on proper style of answering (SFT/DPO) or to teach them learn to reason (GRPO with verifiable reward). STF/DPO is where “stealing” from existing models to save cost of manual labor may happen.

LLM building is nothing that Indian engineers living in India cannot pull off. Don’t worry about Indians who have left. There are plenty in the country as of today.

Then why India does not have foundation models?
———————
It is for the same reason India does not have Google or Facebook of its own.

You need to able to walk before you can run.

There is no protected market to practice your craft in early days. You will get replaced by American service providers as they are cheaper and better every single time. That is not the case with Chinese player. They have a protected market and leadership who treats this skillset as existential due to geopolitics.

So, even if Chinese models are not good in early days they will continue to get funding from their conglomerates as well as provincial governments. Darwinian competition ensures best rise to the top.

Recall DeepSeek took 2 years to get here without much revenue. They were funded by their parent. Also, most of their engineers are not PHDs.

There is nothing that engineers who built Ola/Swiggy/Flipkart cannot build. Remember these services are second to none when you compare them to their Bay Area counterparts. Also , don’t trivialize those services; there is brilliant engineering to make them work at the price points at which they work.

Indian DARPA with 3B USD in funding over 3 years
———————-
What we need is a mentality that treats this skillset as existential. We need a national fund that will fund such teams and the only expected output will be benchmark performance with benchmarks becoming harder every 6 months . No revenue needed to survive for first 3 years.

That money will be loose change for GOI and world’s richest men living in India.

@protosphinx @balajis @vikramchandra @naval

[Sarang Shelke @sarangshelke]

@kaffindia I can’t DM

[Sarang Shelke @sarangshelke]

@kaffindia I am facing simple issue with dishwasher with no resolution for several months now? Official service centre number is not helping - on/off push button switch burnt and you are not able to replace it 🤔

[Sarang Shelke @sarangshelke]

@kaffindia 240209353 this was in feb 2024

[Sarang Shelke @sarangshelke]

@adidas what a flawed defect return policy you have for in store purchases in India. After showing defect product at store, waiting for 2 days and then again returning to store till you approve ?? Which era you operate in. #pathetic

[Sarang Shelke @sarangshelke]

Exciting news! Cadence and Arm have teamed up to launch an ADAS chiplet reference design and software development platform, accelerating time-to-market for software-defined vehicles (SDVs). #Cadence #Arm #ADAS #chiplet #SDV #automotive #innovation

https://t.co/nZf0ocuzY0

[Sarang Shelke @sarangshelke]

@SagasofBharat https://t.co/CjOne9ISog This is run from Nagpur. Also have online short courses. I can vouch for genuine intentions and people running this

[Sarang Shelke @sarangshelke]

Lousy customer service at @TanishqJewelry Aundh Pune. THIRTY minutes to buy gift vouchers. This when they already have all my details in system. #badcustomerservice
#laidbackpune
#TATA needs to improve.

[Sarang Shelke @sarangshelke]

Looks like Cellphone market continues in tailspin.
Chinese cellphone maker Oppo closes its Fabless Chip Division Zeku. https://t.co/9gsxbnFyK2

[Sarang Shelke @sarangshelke]

@sudhirmehtapune @PMCPune If Dutch can return from chaotic car culture to sustainable bike transport and transform their cities, why can’t we. https://t.co/R02plGO2xJ

[Sarang Shelke @sarangshelke]

@sudhirmehtapune @CPPuneCity @ppcr_pune @PuneCityPolice Recent law passed against carpooling apps is regressive. Most developed cities incentives carpooling, This is case of throwing baby with bath water. https://t.co/RJUlYvVUEI

[Sarang Shelke @sarangshelke]

@sudhirmehtapune @ppcr_pune @jagdishmulik @Manojpochat Admire this initiative. In addition to flex - city skyline look ugly with randomly drawn overhead cables both for satellite and fibre optics. Even in smart city showcase areas of Baner, where provision for underground cabling is done, I fail to understand why overhead mess!

[Sarang Shelke @sarangshelke]

Not so subtle.

[Sukhada @appadappajappa]

Ad by Van Gogh Museum for their cafe. Retweet when you get it 😁

[Sarang Shelke @sarangshelke]

@mmpandit It was during thanks giving season of 2002 I had privilege to meet and visit home of Shri PVR Narasimha Rao @homam108 and now after 20 years yog of finally meeting Shri Manish ji @mmpandit 🙏🏼🙏🏼

[Sarang Shelke @sarangshelke]

@MyIndusIndBank what will it take for you to stop giving me 5 calls a day to avail services. Looks like it’s a mistake to hold bank account with you. If these calls don’t stop by tomorrow, I promise all relationships with your bank will be canceled. #harassment

[Sarang Shelke @sarangshelke]

@docbhooshan Reminds of this scene from Shape of Water https://t.co/ZsVodGmUAC 🙂

[Anand Deshpande @anandesh]

We touched a two important revenue milestones
> ₹ 2000 Crores in Quarterly revenue
> US $ 250 Million in the Quarter putting us on a Billion Dollar per year run rate.

Proud of the Persistent Team and #ThankYou all for the support.

0 to 1 Billion USD in 32 years!